Yini Engathenjwa I-Generative AI Ngaphandle Kokungenelela Komuntu?

Isifinyezo esiphezulu

I-Generative Artificial Intelligence (AI) - ubuchwepheshe obuvumela imishini ukuthi idale umbhalo, izithombe, ikhodi, nokunye okwengeziwe - buhlangabezane nokukhula okumangalisayo eminyakeni yamuva nje. Leli phepha elimhlophe linikeza ukubuka konke okufinyelelekayo kokuthi yiziphi i-AI ekhiqizayo engayenza namuhla ngaphandle kokungenelela komuntu, nokuthi yini okulindeleke ukuba iyenze eminyakeni eyishumi ezayo. Sihlola ukusetshenziswa kwayo kukho konke ukubhala, ubuciko, ukubhala amakhodi, isevisi yamakhasimende, ukunakekelwa kwezempilo, imfundo, ukuhlelwa kwempahla, kanye nezezimali, sigqamisa lapho i-AI isebenza ngokuzimele nalapho ukugadwa komuntu kuhlala kubalulekile. Izibonelo zomhlaba wangempela zifakiwe ukuze zibonise kokubili impumelelo nokulinganiselwa. Okutholakele okubalulekile kuhlanganisa:

• Ukutholwa Okusabalele: Ngo-2024, i-65% yezinkampani ezihlolisisiwe zibika njalo zisebenzisa i-AI ekhiqizayo - cishe kabili isabelo kusukela ngonyaka odlule ( Isimo se-AI ekuqaleni kuka-2024 | McKinsey ). Izinhlelo zokusebenza ziqala ukudalwa kokuqukethwe kokumaketha, ama-chatbots okusekelwa kwamakhasimende, ukukhiqiza amakhodi, nokuningi.

• Amakhono Amanje Okuzimela: I-AI yanamuhla ekhiqizayo iphatha ngokuthembekile imisebenzi ehlelekile, ephindaphindayo enokwengamela okuncane. Izibonelo zihlanganisa ukukhiqiza ngokuzenzakalelayo imibiko yezindaba zefomula (isb. izifinyezo zezinzuzo zezinkampani) ( Philana Patterson – ONA Iphrofayela Yomphakathi ), ukukhiqiza izincazelo zomkhiqizo nokubuyekeza okuvelele kumasayithi e-e-commerce, kanye nekhodi yokuqedela ngokuzenzakalela. Kulezi zizinda, i-AI ivamise ukukhuphula izisebenzi zabantu ngokuthatha izintambo zokukhiqiza okuqukethwe okujwayelekile.

• I-Human-in-the-Loop Yemisebenzi Eyinkimbinkimbi: Ukuze uthole imisebenzi eyinkimbinkimbi noma evulekile - njengokubhala ngobuciko, ukuhlaziya okuningiliziwe, noma iseluleko sezokwelapha - ukugadwa komuntu kusadingeka ngokuvamile ukuze kuqinisekiswe ukunemba kweqiniso, ukwahlulela okuhle, kanye nekhwalithi. Ukuthunyelwa okuningi kwe-AI namuhla kusebenzisa imodeli ethi “human-in-the-loop” lapho i-AI ibhala khona okuqukethwe futhi abantu bakubuyekeze.

• Ukuthuthukiswa Kwesikhathi Esiseduze: Phakathi neminyaka engu-5-10 ezayo, i-AI ekhiqizayo kulindeleke ukuthi ithembeke kakhulu futhi izimele . Ukuthuthuka kokunemba kwemodeli nezindlela zokuqapha kungase kuvumele i-AI ukuthi ibambe isabelo esikhulu semisebenzi yokudala neyokwenza izinqumo ngokufaka okuncane komuntu. Ngokwesibonelo, ngo-2030 ochwepheshe babikezela ukuthi i-AI izosingatha iningi lokusebenzisana nezinqumo zesevisi yamakhasimende ngesikhathi sangempela ( Ukuze Ucabange Kabusha I-Shift ku-CX, Abakhangisi Kumele Benze Lezi Zinto Ezingu-2 ), futhi ifilimu enkulu ingase ikhiqizwe ngokuqukethwe okukhiqizwe yi-AI okungu-90% ( I-Generative AI Use Cases for Industries and Enterprises ).

• Ngo-2035: Eminyakeni eyishumi, silindele ukuthi ama-agent e-AI azimele abe yindawo evamile emikhakheni eminingi. Abafundisi be-AI banganikeza imfundo yomuntu siqu ngokwesilinganiso, abasizi be-AI bangase babhale ngokuthembekile izinkontileka zomthetho noma imibiko yezokwelapha ukuze basayine emsebenzini, futhi izinhlelo zokuzishayela (ezisizwa ukulingisa okukhiqizayo) zingase ziqhube imisebenzi yezokuthutha ekugcineni. Kodwa-ke, izindawo ezithile ezibucayi (isb. ukuxilongwa kwezempilo okuphakeme, izinqumo ezingokomthetho zokugcina) cishe zisazodinga ukwahlulela komuntu ngokuphepha nokuziphendulela.

• Izinkathazo Zokuziphatha Nokwethembeka: Njengoba ukuzimela kwe-AI kukhula, kanjalo nokukhathazeka kuyakhula. Izinkinga namuhla zihlanganisa ukubona izinto ezingekho (i-AI eyakha amaqiniso), ukuchema kokuqukethwe okukhiqiziwe, ukuntula obala, kanye nokusetshenziswa kabi okungahle kube khona kwe-disinformation. Ukuqinisekisa ukuthi i-AI ingethenjwa uma isebenza ngaphandle kokwengamela kubalulekile. Inqubekelaphambili iyenziwa - isibonelo, izinhlangano zitshala imali eningi ekunciphiseni ubungozi (ukunemba kokukhuluma, ukuphepha kwe-cybersecurity, izindaba ze-IP) ( Isimo se-AI: Inhlolovo yomhlaba wonke | McKinsey ) - kodwa ukubusa okuqinile nezinhlaka zokuziphatha kuyadingeka.

• Isakhiwo saleli Phepha: Siqala ngesingeniso se-AI ekhiqizayo kanye nomqondo wokusebenzisa ukuzimela uma kuqhathaniswa nokusetshenziswa okugadiwe. Bese, esizindeni esikhulu ngasinye (ukubhala, ubuciko, ukubhala ngekhodi, njll.), sixoxa ngalokho i-AI engayenza ngokwethembeka namuhla uma kuqhathaniswa nalokho okusemkhathizwe. Siphetha ngezinselelo ezihlukanisayo, ukuqagela kwesikhathi esizayo, kanye nezincomo zokusebenzisa ngokunesibopho i-AI ekhiqizayo.

Sekukonke, i-AI ekhiqizayo isivele ifakazele ukuthi iyakwazi ukuphatha uxhaxha olumangalisayo lwemisebenzi ngaphandle kokuqondiswa umuntu njalo. Ngokuqonda imikhawulo yayo yamanje namandla esikhathi esizayo, izinhlangano kanye nomphakathi bangalungiselela kangcono inkathi lapho i-AI ingelona ithuluzi nje, kodwa ingumhlanganyeli ozimele emsebenzini nasekusunguleni.

Isingeniso

I-Artificial Intelligence sekuyisikhathi eside ikwazi ukuhlaziya idatha, kodwa kamuva nje izinhlelo ze-AI zifunde ukudala - ukubhala i-prose, ukuhlanganisa izithombe, isofthiwe yokuhlela, nokuningi. Lawa e-AI akhiqizayo (afana ne-GPT-4 yombhalo noma i-DALL·E yezithombe) aqeqeshelwa amasethi edatha amaningi ukuze akhiqize okuqukethwe kwenoveli ephendula imiyalo. Le mpumelelo iveze igagasi lokuqamba okusha kuzo zonke izimboni. Kodwa-ke, kuphakama umbuzo obalulekile: Yini esingayethemba ngempela ukuthi i-AI izoyenza iyodwa, ngaphandle kokuthi umuntu ahlole kabili ukuphuma kwayo?

Ukuze uphendule lokhu, kubalulekile ukuhlukanisa phakathi okugadwayo nokuzimela - AI:

• I-AI egadwe umuntu isho izimo lapho okuphumayo kwe-AI kubuyekezwa noma kucutshungulwa abantu ngaphambi kokuphothulwa. Isibonelo, intatheli ingase isebenzise umsizi wokubhala we-AI ukuze ibhale isiqephu sendatshana, kodwa umhleli uyasihlela futhi asigunyaze.

• I-Autonomous AI (AI ngaphandle kokungenelela komuntu) ibhekisela kumasistimu e-AI enza imisebenzi noma akhiqiza okuqukethwe okungena ukusetshenziswa ngokuqondile ngokuhlelwa okuncane noma okungekho nhlobo komuntu. Isibonelo i-chatbot ezenzakalelayo exazulula umbuzo wekhasimende ngaphandle komenzeli ongumuntu, noma isitolo sezindaba esishicilela ngokuzenzakalela isifinyezo semiphumela yezemidlalo esikhiqizwe i-AI.

I-Generative AI isivele isetshenziswa kuzo zombili izindlela. Ngo-2023-2025, ukutholwa kuye kwanda kakhulu , nezinhlangano zizama ngokulangazela. Inhlolovo eyodwa yomhlaba wonke ngo-2024 yathola ukuthi izinkampani ezingu-65% zisebenzisa njalo i-AI ekhiqizayo, kusukela cishe kokukodwa kokuthathu ngonyaka odlule ( Isimo se-AI ekuqaleni kuka-2024 | McKinsey ). Abantu ngabanye, nabo, bamukele amathuluzi afana ne-ChatGPT - okulinganiselwa ukuthi ama-79% ochwepheshe okungenani abe nokuchayeka okuthile ku-AI ekhiqizayo maphakathi no-2023 ( Isimo se-AI ngo-2023: unyaka wokuphuma kwe-Generative AI | McKinsey ). Lokhu kuthathwa ngokushesha kuqhutshwa isithembiso sokusebenza kahle kanye nezinzuzo zokusungula. Nokho kusasele “izinsuku zakuqala,” futhi izinkampani eziningi zisabumba izinqubomgomo zokuthi isetshenziswa kanjani i-AI ngokuzibophezela ( Isimo se-AI ngo-2023: Unyaka wokuphuma kwe-Generative AI | McKinsey ).

Kungani ukuzimela kubalulekile: Ukuvumela i-AI ukuthi isebenze ngaphandle kokuqondisa komuntu kungavula izinzuzo ezinkulu zokusebenza kahle - ukuzenzela imisebenzi eyisicefe ngokuphelele - kodwa futhi kuphakamisa izigxobo zokwethembeka. I-ejenti ye-AI ezimele kufanele ilungise izinto (noma yazi imikhawulo yayo) ngoba kungase kungabikho muntu ngesikhathi sangempela wokubamba amaphutha. Eminye imisebenzi iboleka kulokhu ngaphezu kweminye. Ngokuvamile, i-AI isebenza kahle kakhulu ngokuzimele lapho:

• Umsebenzi unesakhiwo esicacile noma iphethini (isb. ukukhiqiza imibiko yesimiso ngedatha).

• Amaphutha anengozi ephansi noma abekezelelwa kalula (isb. ukukhiqizwa kwesithombe okungalahlwa uma kungagculisi, uma kuqhathaniswa nokuxilongwa kwezokwelapha).

• Kunedatha eyanele yokuqeqeshwa ehlanganisa izimo, ngakho-ke okukhiphayo kwe-AI kusekelwe ezibonelweni zangempela (ukunciphisa ukuqagela).

Ngokuphambene, imisebenzi eneziphetho ezivulekile , ephezulu , noma edinga ukwahlulela okucashile ayifanelekeli ukwengamela iqanda namuhla.

Ezigabeni ezilandelayo, sihlola inqwaba yezinkambu ukuze sibone ukuthi i-generative AI yenzani manje nokuthi yini elandelayo. Sizobheka izibonelo ezibambekayo - kusukela ezindabeni zezindaba ezibhalwe nge-AI kanye nomsebenzi wobuciko okhiqizwe yi-AI, kuya kubasizi bokubhala amakhodi nama-ejenti wesevisi yamakhasimende abonakalayo - sigqamisa ukuthi yimiphi imisebenzi engenziwa i-AI ekugcineni futhi esadinga umuntu owaziyo. Esizindeni ngasinye, sihlukanisa ngokucacile amandla amanje (circa 2025) kusukela ekuqageleni okungokoqobo kwalokho okungathembeka ngo-2035.

Ngokwenza imephu yamanje nekusasa le-AI ezimele kuzo zonke izizinda, sihlose ukunikeza abafundi ukuqonda okunokulinganisela: hhayi ukweqisa i-AI njengengenaphutha ngomlingo, noma ukudayisa kancane amakhono ayo angempela nakhulayo. Ngalesi sisekelo, sibe sesixoxa ngezinselele ezinkulu ekuthembeni i-AI ngaphandle kokugadwa, okuhlanganisa ukucatshangelwa kwezimiso zokuziphatha nokulawulwa kobungozi, ngaphambi kokuphetha ngezinto ezibalulekile ezithathwayo.

I-Generative AI ekubhaleni nasekudaleni okuqukethwe

Esinye sezizinda zokuqala lapho i-AI ekhiqizayo yenza ukuchaphazeka kwaba ukukhiqizwa kombhalo. Amamodeli olimi amakhulu angakhiqiza yonke into kusukela kuzihloko zezindaba kanye nekhophi yokumaketha kuya kokuthunyelwe kwenkundla yezokuxhumana kanye nezifinyezo zamadokhumenti. Kodwa kungakanani kwalokhu kubhala okungenziwa ngaphandle komhleli ongumuntu?

Amandla Amanje (2025): I-AI Njengombhali Ozenzakalelayo Wokuqukethwe Komjikelezo

Namuhla, i-AI ekhiqizayo iphatha ngokuthembekile imisebenzi ehlukahlukene yokubhala ngokungenelela okuncane noma okungekho komuntu. Isibonelo esiyinhloko esobuntatheli: I-Associated Press iminyaka eminingi isebenzisa okuzenzakalelayo ukuze ikhiqize izinkulungwane zemibiko yemali etholwa yinkampani ikota ngayinye ngokuqondile ngokuphakelayo kwedatha yezezimali ( Philana Patterson - ONA Iphrofayela Yomphakathi ). Lezi ziqephu zezindaba ezimfushane zilandela isifanekiso (isb., “Inkampani X ibike amaholo ka-Y, akhuphuke ngo-Z%....”) futhi i-AI (isebenzisa isofthiwe yokukhiqiza ulimi lwemvelo) ingagcwalisa izinombolo nezwi ngokushesha kunanoma yimuphi umuntu. Uhlelo lwe-AP lushicilela le mibiko ngokuzenzakalelayo, lwandise ukufakwa kwayo ngendlela emangazayo (ngaphezu kwezindaba ezingu-3,000 ngekota ngayinye) ngaphandle kokudinga ababhali abangabantu ( Izindaba zeholo elizenzakalelayo ziyaphindaphindeka | The Associated Press ).

Ubuntatheli bezemidlalo buye bandiswa ngendlela efanayo: Amasistimu e-AI angathatha izibalo zegeyimu yezemidlalo futhi akhiqize izindaba eziphindaphindayo. Ngenxa yokuthi lezi zizinda ziqhutshwa yidatha futhi zihlelekile, amaphutha awavamile uma nje idatha ilungile. Kulezi zimo, sibona ukuzimela kweqiniso - i-AI iyabhala futhi okuqukethwe kushicilelwa ngaso leso sikhathi.

Amabhizinisi asebenzisa i-AI ekhiqizayo ukuze abhale izincazelo zomkhiqizo, izincwadi zezindaba ze-imeyili, nokunye okuqukethwe kokumaketha. Isibonelo, isiqhwaga se-e-commerce i-Amazon manje sesisebenzisa i-AI ukufingqa ukubuyekezwa kwamakhasimende ngemikhiqizo. I-AI iskena umbhalo wezibuyekezo eziningi zomuntu ngamunye futhi ikhiqize isigaba esivelele esifingqiwe salokho abantu abakuthandayo noma abangakuthandi mayelana nento, ebese iboniswa ekhasini lomkhiqizo ngaphandle kokuhlela mathupha ( I-Amazon ithuthukisa isipiliyoni sokubuyekezwa kwamakhasimende nge-AI ). Ngezansi kunomfanekiso walesi sici esisetshenziswe kuhlelo lokusebenza lweselula lwe-Amazon, lapho isigaba esithi “Abathengi bathi” sikhiqizwe ngokuphelele yi-AI kusuka kudatha yokubuyekeza:

( I-Amazon ithuthukisa ulwazi lokubuyekezwa kwamakhasimende nge-AI ) Isifinyezo sokubuyekezwa esikhiqizwe yi-AI ekhasini lomkhiqizo we-e-commerce. Uhlelo lwe-Amazon lufingqa amaphuzu avamile avela ekubuyekezweni kwabasebenzisi (isb., ukusetshenziswa kalula, ukusebenza) kusigaba esifushane, esiboniswa abathengi njengokuthi “I-AI-ekhiqizwe kusukela kumbhalo wokubuyekezwa kwamakhasimende.”

Izimo ezinjalo zokusetshenziswa zibonisa ukuthi lapho okuqukethwe kulandela iphethini ebikezelwayo noma kuhlanganiswa kudatha ekhona, i-AI ngokuvamile ingakwazi ukuyiphatha iyodwa . Ezinye izibonelo zamanje zifaka:

• Isimo Sezulu Nezibuyekezo Zethrafikhi: Izitolo zemidiya ezisebenzisa i-AI ukuze zihlanganise imibiko yesimo sezulu yansuku zonke noma izaziso zethrafikhi ngokusekelwe kudatha yezinzwa.

• Imibiko Yezezimali: Amafemu akhiqiza izifinyezo zezezimali eziqondile (imiphumela yekota, ukwaziswa ngezimakethe zamasheya) ngokuzenzakalelayo. Kusukela ngo-2014, i-Bloomberg nezinye izitolo zezindaba zisebenzise i-AI ukusiza ekubhaleni ama-blurb wezindaba ngemali etholwayo yenkampani - inqubo esebenza ngokuzenzakalelayo uma idatha isitholiwe ( 'izintatheli zerobhothi' ze-AP zibhala izindaba zazo manje | The Verge ) ( Intatheli yase-Wyoming ibanjwe isebenzisa i-AI ukuze idale izingcaphuno ezingelona iqiniso, izindaba ).

• Ukuhumusha Nokulotshwa: Amasevisi okuloba manje asebenzisa i-AI ukuze akhiqize imibhalo yomhlangano noma amagama-ncazo ngaphandle kwabashicileli abangabantu. Nakuba ingakhiqizi ngomqondo wokudala, le misebenzi yolimi isebenza ngokuzenzakalelayo ngokunemba okuphezulu komsindo ocacile.

• Isizukulwane Esisalungiswa: Ochwepheshe abaningi basebenzisa amathuluzi afana ne-ChatGPT ukuze babhale ama-imeyili noma izinguqulo zokuqala zamadokhumenti, ngezikhathi ezithile bewathumela ngokuhlelwa okuncane noma okungenasiphelo uma okuqukethwe kunobungozi obuncane.

Kodwa-ke, nge-prose eyinkimbinkimbi, ukubhekwa komuntu kuhlala kuyinto evamile ku-2025 . Izinhlangano zezindaba azivamile ukushicilela izindatshana eziphenyayo noma ezihlaziyayo zisuka ku-AI - abahleli bazohlola iqiniso futhi balungise okusalungiswa okulotshwe nge-AI. I-AI ingakwazi ukulingisa isitayela nesakhiwo kahle kodwa ingase yethule amaphutha amaqiniso (ngokuvamile abizwa ngokuthi “ama-hallucinations”) noma imishwana engacacile umuntu okudingeka ayibambe. Isibonelo, iphephandaba laseJalimane i-Express yethula "uzakwethu wedijithali" we-AI okuthiwa u-Klara ukusiza ukubhala izingcezu zezindaba zokuqala. U-Klara angabhala kahle imibiko yezemidlalo futhi abhale nezihloko zezindaba eziheha ukufunda, okunikela ku-11% wezindatshana ze-Express - kodwa abahleli abangabantu basabuyekeza zonke izingxenye ukuze bathole ukunemba nobuqotho bezintatheli, ikakhulukazi ezindabeni eziyinkimbinkimbi ( Izindlela Eziyi-12 Izintatheli Ezisebenzisa Amathuluzi E-AI Egumbini Lezindaba - Twipe ). Lokhu kubambisana kwabantu ne-AI kuvamile namuhla: I-AI isingatha ukuphakanyiswa okusindayo kokukhiqiza umbhalo, futhi abantu bakhetha futhi balungise njengoba kudingeka.

I-Outlook ka-2030-2035: Ibheke Ekubhaleni Okuzenzakalelayo Okwethenjwayo

Kule minyaka eyishumi ezayo, silindele ukuthi i-AI ekhiqizayo ithembeke kakhulu ekukhiqizeni umbhalo osezingeni eliphezulu, oyiqiniso, ozonweba ububanzi bemisebenzi yokubhala engakwazi ukuwenza ngokuzenzakalelayo. Amathrendi amaningana asekela lokhu:

• Ukunemba Okuthuthukisiwe: Ucwaningo oluqhubekayo lwehlisa ngokushesha ukuthambekela kwe-AI ukukhiqiza ulwazi olungamanga noma olungabalulekile. Ngo-2030, amamodeli olimi athuthukisiwe anokuqeqeshwa okungcono (okuhlanganisa namasu okuqinisekisa amaqiniso ngokumelene nesizindalwazi ngesikhathi sangempela) angafinyelela ezingeni eliseduze lokuhlola iqiniso ngaphakathi. Lokhu kusho ukuthi i-AI ingase ibhale isiqephu sendatshana sezindaba esigcwele ngezingcaphuno ezifanele kanye nezibalo ezikhishwe kumthombo ngokuzenzekelayo, okudinga ukuhlelwa okuncane.

• I-Domain-Specific AIs: Sizobona amamodeli akhiqizayo akhethekile ashunwe kahle emikhakheni ethile (ukubhala kwezomthetho, kwezokwelapha, kobuchwepheshe). Imodeli yezomthetho ye-AI yango-2030 ingase ibhale ngokuthembekile izinkontileka ezijwayelekile noma ifinyeze umthetho wamacala - imisebenzi ehlelwe ngendlela ehlelekile kodwa okwamanje efuna isikhathi sommeli. Uma i-AI iqeqeshwe kumadokhumenti omthetho aqinisekisiwe, okusalungiswa kwayo kungase kuthembeke ngokwanele kangangokuthi ummeli unikeza nje ukubheka okusheshayo kokugcina.

• Isitayela Semvelo Nokubumbana: Amamodeli aba ngcono ekulondolozeni umongo kumadokhumenti amade, okuholela kokuqukethwe okuhambisanayo kanye nephuzu lefomu elide. Ngo-2035, kuyathandeka ukuthi i-AI ingabhala uhlaka lokuqala oluhle lwencwadi engelona iqiniso noma imanuwali yobuchwepheshe iyodwa, nabantu ikakhulukazi endimeni yokweluleka (ukubeka imigomo noma ukuhlinzeka ngolwazi olukhethekile).

Lokhu kungase kubukeke kanjani ekusebenzeni? Ubuntatheli obujwayelekile bungase bucishe bube ngokuzenzakalela kumabhithi athile. Singabona inkampani yezindaba ngo-2030 inesistimu ye-AI ibhala inguqulo yokuqala yawo wonke umbiko wemali engenayo, indaba yezemidlalo, noma isibuyekezo semiphumela yokhetho, umhleli ethatha amasampula ambalwa kuphela ukuze aqinisekise ikhwalithi. Ngempela, ochwepheshe babikezela ukuthi isabelo esikhula njalo sokuqukethwe ku-inthanethi sizokwenziwa ngomshini - ukubikezela okukodwa okunesibindi kwabahlaziyi bemboni kuphakamise ukuthi kufikela ku-90% wokuqukethwe ku-inthanethi kungase kube yi-AI-ekhiqizwe ngo-2026 ( Ngo-2026, Okuqukethwe Ku-inthanethi Okukhiqizwe Abantu Okungebona Abantu Kuzokwedlula Kakhulu Inani Lokuqukethwe Okukhiqizwa Abantu - I-OODAloop iyinkulumompikiswano). Ngisho nomphumela oqinile ungasho ukuthi maphakathi nawo-2030, iningi lama-athikili ewebhu, ikhophi yomkhiqizo, futhi mhlawumbe nokuphakelayo kwezindaba komuntu siqu kugunyazwe yi-AI.

Kwezokumaketha nokuxhumana kwezinkampani , i-AI ekhiqizayo cishe izophathiswa ukuthi iqhube imikhankaso yonke ngokuzimele. Ingase ikhiqize futhi ithumele ama-imeyili okumaketha omuntu siqu, okuthunyelwe kwenkundla yezokuxhumana, kanye nokuhluka kwamakhophi esikhangiso, ishintsha njalo umlayezo ngokusekelwe ekuphenduleni kwamakhasimende - konke ngaphandle komkopishi womuntu ku-loop. Abahlaziyi be-Gartner baphrojekthi ukuthi ngo-2025, okungenani u-30% wemilayezo yokumaketha ephumayo yamabhizinisi amakhulu izobe ikhiqizwa i-AI ( Generative AI Use Cases for Industries and Enterprises ), futhi leli phesenti lizokhuphuka kuphela ngo-2030.

Nokho, kubalulekile ukuqaphela ukuthi ubuhlakani bomuntu nokwahlulela kusazodlala indima, ikakhulukazi kokuqukethwe okuphakeme . Ngo-2035, i-AI ingase isingathe ukukhishwa kwabezindaba noma okuthunyelwe kwebhulogi iyodwa, kodwa kubuntatheli obuphenyayo obubandakanya ukuziphendulela noma izihloko ezibucayi, izitolo zezindaba zisengagcizelela ekubhekeni abantu. Ikusasa lingase lilethe indlela enezigaba: I-AI ikhiqiza ngokuzenzakalelayo inqwaba yokuqukethwe kwansuku zonke, kuyilapho abantu begxile ekuhleleni nasekukhiqizeni izingcezu zamasu noma ezibucayi. Empeleni, ulayini walokho okubalulekile “njengokujwayelekile” uzokhula njengoba ubuhlakani be-AI bukhula.

Ukwengeza, izinhlobo ezintsha zokuqukethwe ezifana nezindaba ezisebenzisanayo ezikhiqizwe yi-AI noma imibiko yomuntu siqu ingase ivele. Isibonelo, umbiko wonyaka wenkampani ungenziwa ngezitayela eziningi yi-AI - kafushane kubaphathi, inguqulo elandisayo yabasebenzi, inguqulo enothile yedatha yabahlaziyi - ngayinye idalwe ngokuzenzakalelayo kusuka kudatha efanayo eyisisekelo. Kwezemfundo, izincwadi zokufunda zingabhalwa ngamandla yi-AI ukuze zivumelane namazinga okufunda ahlukene. Lezi zinhlelo zokusebenza zingase zizimele kodwa zisekelwe ulwazi oluqinisekisiwe.

I-trajectory ekubhaleni iphakamisa ukuthi maphakathi nawo-2030, i-AI izoba umbhali ophumelelayo . Ukhiye wokusebenza ngokuzimela ngempela uzosungula ukwethenjwa emiphumeleni yakhona. Uma i-AI ingabonisa ngokungaguquki ukunemba kweqiniso, ikhwalithi yesitayela, nokuqondanisa nezindinganiso zokuziphatha, isidingo sokubuyekezwa komuntu ngomugqa ngamunye sizoncipha. Izigaba zaleli phepha elimhlophe ngokwalo, ngo-2035, zingahle zibhalwe umcwaningi we-AI ngaphandle kokudinga umhleli - ithemba esinethemba elikhulu ngalo, inqobo nje uma kunezivikelo ezifanele.

I-Generative AI ku-Visual Arts and Design

Ikhono le-Generative AI lokudala izithombe nomsebenzi wobuciko lithathe umcabango womphakathi, kusukela emiqhudelwaneni yobuciko ekhiqizwe yi-AI kuya kumavidiyo mbumbulu angenakuhlukaniswa nezithombe zangempela. Ezizindeni ezibukwayo, amamodeli e-AI afana namanethiwekhi ezitha ezikhiqizayo (ama-GAN) namamodeli okusabalalisa (isb. I-Stable Diffusion, Midjourney) angakhiqiza izithombe zoqobo ngokusekelwe ekwazisweni kombhalo. Ngakho-ke, ingabe i-AI manje ingasebenza njengeciko elizimele noma umklami?

Amakhono Amanje (2025): I-AI Njengomsizi Wokudala

Kusukela ngo-2025, amamodeli akhiqizayo anekhono ekudaleni izithombe uma zidingeka ngokwethembeka okuhlaba umxhwele. Abasebenzisi bangacela isithombe se-AI ukuthi sidwebe “idolobha lasendulo ekushoneni kwelanga ngesitayela sikaVan Gogh” futhi bathole isithombe sobuciko obukholisayo ngemizuzwana. Lokhu kuholele ekusetshenzisweni okubanzi kwe-AI ekwakhiweni kwengcaca, ukumaketha, nokuzijabulisa kobuciko bomqondo, ama-prototypes, ngisho nokubonwa kokugcina kwezinye izimo. Ngokuphawulekayo:

• I-Graphic Design & Stock Images: Izinkampani zikhiqiza imidwebo yewebhusayithi, imidwebo, noma izithombe zesitoko nge-AI, okunciphisa isidingo sokuthumela zonke izingcezu ezivela kumdwebi. Amathimba amaningi okuthengisa asebenzisa amathuluzi e-AI ukukhiqiza okuhlukile kwezikhangiso noma izithombe zomkhiqizo ukuze ahlole ukuthi yini ekhanga abathengi.

• Ubuciko Nemifanekiso: Amaciko angawodwana ahlanganyela ne-AI ukuze bacabangele imibono noma bagcwalise imininingwane. Isibonelo, umdwebi wezithombe angase asebenzise i-AI ukuze enze indawo engemuva, abe eseyihlanganisa nezinhlamvu zabo ezidwetshwe umuntu. Abanye abadali bamabhuku amahlaya bazame ngamaphaneli akhiqizwe yi-AI noma imibala.

• Imidiya Nokuzijabulisa: Ubuciko obukhiqizwe nge-AI buvele kumakhava omagazini namakhava ezincwadi. Isenzakalo esidumile kwakuyikhava yango-August 2022 ye-Cosmopolitan eyayibonisa usomkhathi - okubikwa ukuthi isithombe sokuqala sekhava yemagazini esakhiwe i-AI (i-OpenAI's DALL·E) njengoba iqondiswa umqondisi wezobuciko. Nakuba lokhu kwakuhilela ukukhuthazwa nokukhetha komuntu, umsebenzi wobuciko wangempela wawunikezwa ngomshini.

Okubalulekile, okuningi kwalokhu kusetshenziswa kwamanje kusabandakanya ukunakekelwa komuntu kanye nokuphindaphinda . I-AI ingakhafula inqwaba yezithombe, futhi umuntu ukhetha okungcono kakhulu futhi ngokunokwenzeka akuthinte. Ngalowo mqondo, i-AI isebenza ngokuzimela ukuze ikhiqize izinketho, kodwa abantu baqondisa indlela yokudala futhi benza ukukhetha kokugcina. Kuthembekile ekukhiqizeni okuqukethwe okuningi ngokushesha, kodwa akuqinisekisiwe ukuhlangabezana nazo zonke izimfuneko ekuzameni kokuqala. Izinkinga ezifana nemininingwane engalungile (isb. Ukudweba izandla nge-AI ngenombolo yeminwe engalungile, into eyaziwayo) noma imiphumela engahlosiwe isho ukuthi umqondisi wezobuciko ongumuntu ngokuvamile udinga ukugada ikhwalithi yokuphumayo.

Kukhona, nokho, izizinda lapho i-AI isondela ekuzilawuleni okuphelele:

• Idizayini Ekhiqizayo: Emikhakheni efana nezakhiwo kanye nokwakhiwa komkhiqizo, amathuluzi e-AI angakwazi ukudala ngokuzenzakalelayo ama-prototypes edizayini ahlangabezana nezingqinamba ezithile. Isibonelo, uma kubhekwa ubukhulu obufunwayo kanye nemisebenzi yefenisha, i-algorithm ekhiqizayo ingase ikhiphe imiklamo eminingana esebenzayo (eminye engajwayelekile) ngaphandle kokungenelela komuntu ngale kwemininingwane yokuqala. Le miklamo ingase isetshenziswe ngokuqondile noma icwengwe abantu. Ngokufanayo, kwezobunjiniyela, i-AI ekhiqizayo ingakwazi ukuklama izingxenye (ake sithi, ingxenye yendiza) elungiselelwe isisindo namandla, ikhiqize umumo wenoveli umuntu okungenzeka ukuthi akazange awakhulelwe.

• Impahla Yegeyimu yevidiyo: I-AI ingakhiqiza ukuthungwa, amamodeli e-3D, noma ngisho namaleveli aphelele emidlalo yevidiyo ngokuzenzakalela. Onjiniyela basebenzisa lokhu ukusheshisa ukudalwa kokuqukethwe. Amanye amageyimu e-indie aseqalile ukuhlanganisa umsebenzi wobuciko owenziwe ngokwenqubo kanye nengxoxo (ngamamodeli olimi) ukuze adale umhlaba omkhulu, oguquguqukayo wamageyimu onezimpahla ezincane ezidalwe abantu.

• Ukugqwayiza Nevidiyo (Esafufusa): Nakuba ikhulile kancane kunezithombe ezimile, i-AI ekhiqizayo yevidiyo iyathuthuka. I-AI isingakwazi kakade ukukhiqiza iziqeshana zevidiyo ezimfushane noma izithombe ezinyakazayo ezivela ekwazisweni, nakuba ikhwalithi ingahambisani. Ubuchwepheshe be-Deepfake - obukhiqizayo - bungakhiqiza ukushintshaniswa kobuso okungokoqobo noma ama-clones wezwi. Esimweni esilawulwayo, isitudiyo singasebenzisa i-AI ukuze sikhiqize indawo engemuva noma ukugqwayiza kwesixuku ngokuzenzakalela.

Ngokuphawulekayo, u-Gartner ubikezele ukuthi ngo-2030, sizobona ifilimu enkulu ye-blockbuster ene-90% yokuqukethwe okukhiqizwe yi-AI (kusuka kuskripthi kuya kokubonwayo) ( I-Generative AI Use Cases for Industries and Enterprises ). Kusukela ngo-2025, asikabi lapho - i-AI ayikwazi ukwenza ngokuzimela ifilimu yobude besici. Kodwa izingcezu zaleyo puzzle ziyathuthuka: ukukhiqizwa kombhalo (umbhalo we-AI), ukukhiqizwa komlingiswa nesigcawu (isithombe/ividiyo i-AI), ukwenza izwi (ama-clone wezwi we-AI), nosizo lokuhlela (i-AI isingasiza kakade ngokusikeka noshintsho).

I-Outlook ka-2030-2035: Imidiya Ekhiqizwe Nge-AI Esikalini

Uma ubheka phambili, indima ye-AI ekhiqizayo kwezobuciko bokubukwayo nokuklama isilungele ukwanda kakhulu. Ngo-2035, silindele ukuthi i-AI izoba ngumdali wokuqukethwe oyinhloko kumidiya eminingi ebonakalayo, evame ukusebenza ngokufaka okuncane komuntu ngaphezu kokuqondiswa kokuqala. Okunye okulindelekile:

• Amafilimu Namavidiyo Akhiqizwe Nge-AI Ngokugcwele: Eminyakeni eyishumi ezayo, kungenzeka ukuthi sizobona amamuvi okuqala noma uchungechunge olukhiqizwe kakhulu nge-AI. Abantu bangase banikeze ukuqondiswa kwezinga eliphezulu (isb. uhlaka lweskripthi noma isitayela esifiswayo) futhi i-AI izonikeza izigcawu, idale ukufana nomlingisi, futhi iphilise yonke into. Ukuhlolwa kwangaphambi kwesikhathi kumafilimu amafushane kungenzeka phakathi neminyaka embalwa, ngemizamo yobude besici ngawo-2030. Lawa mafilimu e-AI angase aqale i-niche (ukugqwayiza kokuhlola, njll.) kodwa angase abe yinsakavukela njengoba ikhwalithi ithuthuka. Isibikezelo sefilimu sika-Gartner sika-90% ngo-2030 ( I-Generative AI Use Cases for Industries and Enterprises ), kuyilapho isifiso sokuvelela, igcizelela inkolelo yemboni yokuthi ukudalwa kokuqukethwe kwe-AI kuzoba yinkimbinkimbi ngokwanele ukuthwala umthwalo omningi ekwenzeni amafilimu.

• I-Design Automation: Emikhakheni efana nemfashini noma izakhiwo, i-AI ekhiqizayo cishe izosetshenziselwa ukubhala ngokuzenzakalelayo amakhulu emibono yokuklama ngokusekelwe kumapharamitha afana “nezindleko, izinto zokwakha, isitayela X”, okushiya abantu bekhetha umklamo wokugcina. Lokhu kushintsha amandla amanje: esikhundleni sokuthi abaklami badale kusukela ekuqaleni futhi mhlawumbe basebenzise i-AI ukuze bathole ugqozi, abaklami bakusasa bangase basebenze njengabadidiyeli, bakhethe umklamo ongcono kakhulu okhiqizwe yi-AI futhi mhlawumbe bawulungise. Ngo-2035, umakhi angase afake izidingo zebhilidi futhi athole amapulani aphelele njengeziphakamiso ezivela ku-AI (konke kuzwakala kahle ngokwesakhiwo, ngezilokotho zemithetho eshumekiwe yobunjiniyela).

• Ukudalwa Kokuqukethwe Okungokwezifiso: Singabona i-AI idala okubonwayo ngokushesha kubasebenzisi abangabodwana. Cabanga ngegeyimu yevidiyo noma okuhlangenwe nakho okungokoqobo okubonakalayo ngo-2035 lapho indawo nezinhlamvu zijwayelana nokuthandwayo komdlali, okukhiqizwa ngesikhathi sangempela yi-AI. Noma imicu yamahlaya yomuntu siqu ekhiqizwe ngokusekelwe osukwini lomsebenzisi - i-AI "yamahlaya yedayari yansuku zonke" eshintshayo eguqula iphephabhuku lakho libe yimifanekiso ngokuzenzakalelayo njalo kusihlwa.

• I-Multimodal Creativity: Amasistimu e-Generative AI aya ngokuya enziwa ngezinto eziningi - okusho ukuthi angakwazi ukuphatha umbhalo, izithombe, umsindo, njll. ndawonye. Ngokuhlanganisa lokhu, i-AI ingathatha ukwaziswa okulula njengokuthi “Ngenze umkhankaso wokumaketha womkhiqizo X” futhi ingakhiqizi nje ikhophi ebhaliwe, kodwa ihluzo elihambisanayo, mhlawumbe neziqeshana zevidiyo ezimfushane ezikhangisayo, zonke ezingaguquki ngesitayela. Lolu hlobo lokuchofoza okukodwa kokuqukethwe luyisevisi okungenzeka ekuqaleni kwawo-2030.

Ingabe i-AI izongena esikhundleni sabaculi abangabantu ? Lo mbuzo uvame ukuvela. Kungenzeka ukuthi i-AI izothatha umsebenzi omningi wokukhiqiza (ikakhulukazi ubuciko obuphindaphindayo noma obushintsha ngokushesha obudingekayo ebhizinisini), kodwa ubuciko bomuntu buyohlala bungokwangempela kanye nokusungula izinto ezintsha. Ngo-2035, i-AI ezimele ingase idwebe isithombe ngendlela enokwethenjelwa ngesitayela seciko elidumile - kodwa ukudala esisha noma ubuciko obunamasiko ajulile kusengase kube inhlanhla yomuntu (okungenzeka kube ne-AI njengomhlanganyeli). Sibona kusengaphambili ikusasa lapho abaculi abangabantu besebenza khona kanye “nabaculi ababambisene” be-AI abazimele. Umuntu angase athumele i-AI yomuntu siqu ukuthi iqhubeke ikhiqize ubuciko begalari yedijithali ekhaya lomuntu, isibonelo, ukuhlinzeka nge-ambience yokudala ehlala ishintsha.

Ngokombono wokwethenjelwa, i-AI ekhiqizayo ebonakalayo inendlela elula yokuzimela kunombhalo ngezindlela ezithile: isithombe singaba “sihle ngokwanele” noma singaphelele, kanti iphutha eliyiqiniso embhalweni liyinkinga kakhulu. Ngakho-ke, sesivele sibona ukutholwa okunengozi ephansi - uma idizayini ekhiqizwe yi-AI iyimbi noma ingalungile, awusebenzisi nje, kodwa ayizilimazi yona ngokwayo. Lokhu kusho ukuthi ngeminyaka yawo-2030, izinkampani zingase zikhululeke ukuvumela i-AI ukuthi ikhiphe imiklamo ingagadiwe futhi ifaka abantu kuphela lapho kudingeka inoveli yangempela noma eyingozi.

Kafushane, ngo-2035 i-AI ekhiqizayo kulindeleke ukuthi ibe ngumdali wokuqukethwe onamandla ezibonwayo, okungenzeka kube nesibopho engxenyeni enkulu yezithombe nabezindaba ezisizungezile. Izokhiqiza ngokuthembekile okuqukethwe kwezokuzijabulisa, idizayini, nokuxhumana kwansuku zonke. Iciko elizimele liseduze - noma ngabe i-AI ibonwa njengokudala noma ithuluzi elihlakaniphe kakhulu kuyinkulumo-mpikiswano ezovela njengoba imiphumela yayo ingabonakali kokwenziwe umuntu.

I-Generative AI ekuthuthukisweni kweSoftware (Coding)

Ukuthuthukiswa kwesofthiwe kungase kubonakale kuwumsebenzi wokuhlaziya kakhulu, kodwa futhi kunesici sokudala - ikhodi yokubhala idala umbhalo ngolimi oluhlelekile. I-AI yesimanjemanje, ikakhulukazi amamodeli ezilimi ezinkulu, afakazele ukuthi anekhono ekubhaleni amakhodi. Amathuluzi afana ne-GitHub Copilot, i-Amazon CodeWhisperer, namanye asebenza njengabahleli bezinhlelo be-AI, ephakamisa amazwibela ekhodi noma imisebenzi yonke njengohlobo lonjiniyela. Lokhu kungafinyelela kude kangakanani ekuhleleni okuzimele?

Amakhono Amanje (2025): I-AI Njengomshayeli Wokushayela Ngekhodi

Ngo-2025, abakhiqizi bekhodi ye-AI sebejwayelekile ekusebenzeni konjiniyela abaningi. Lawa mathuluzi angaqedela ngokuzenzakalela imigqa yekhodi, enze i-boilerplate (njengemisebenzi evamile noma ukuhlola), futhi abhale izinhlelo ezilula ezinikezwe incazelo yolimi lwemvelo. Okubalulekile, nokho, zisebenza ngaphansi kokugadwa unjiniyela - unjiniyela ubuyekeza futhi ahlanganise iziphakamiso ze-AI.

Amanye amaqiniso nezibalo zamanje:

• Ngaphezu kwesigamu sabathuthukisi abangochwepheshe basebamukele abasizi bokufaka amakhodi be-AI ngasekupheleni kuka-2023 ( Ukubhala ngekhodi ku-Copilot: 2023 Idatha Iphakamisa Ukucindezela Okuya Phansi Ngekhwalithi Yekhodi (okuhlanganisa nokuqagela kuka-2024) - GitClear ), okubonisa ukutholwa ngokushesha. I-GitHub Copilot, elinye lamathuluzi okuqala atholakala kabanzi, kubikwe ukuthi ikhiqize ngokwesilinganiso u-30-40% wekhodi kumaphrojekthi lapho isetshenziswa khona ( Ukubhala ngekhodi akuseyona i-MOAT. U-46% wamakhodi ku-GitHub usuvele ... ). Lokhu kusho ukuthi i-AI isivele ibhala izingxenye ezibalulekile zekhodi, nakuba umuntu eyiqondisa futhi eyiqinisekisa.

• Lawa mathuluzi e-AI ahamba phambili emisebenzini efana nokubhala ikhodi ephindaphindwayo (isb, amakilasi emodeli yedatha, izindlela ze-getter/setter), ukuguqula ulimi oluthile lokuhlela lube kolunye, noma ukukhiqiza ama-algorithms aqondile afana nezibonelo zokuqeqesha. Isibonelo, unjiniyela angaphawula "// umsebenzi ukuze ahlele uhlu lwabasebenzisi ngamagama" futhi i-AI izokhiqiza umsebenzi wokuhlunga ofanele cishe ngaso leso sikhathi.

• Baphinde basize ekulungiseni iziphazamisi kanye nencazelo : onjiniyela banganamathisela umlayezo wephutha futhi i-AI ingase iphakamise ukulungisa, noma ibuze ukuthi "Lenzani le khodi?" futhi uthole incazelo yolimi lwemvelo. Lokhu kuzimele ngomqondo othile (i-AI ingakwazi ukuhlonza izinkinga ngokwayo), kodwa umuntu unquma ukuthi uzokusebenzisa yini ukulungisa.

• Okubalulekile, abasizi bekhodi be-AI bamanje abanaphutha. Bangaphakamisa ikhodi engavikelekile, noma ikhodi ecishe ixazulule inkinga kodwa eneziphazamisi ezicashile. Ngakho-ke, umkhuba ongcono kakhulu namuhla ukugcina umuntu ekwazini - unjiniyela uhlola futhi alungise ikhodi ebhalwe nge-AI njengoba bebengenza ikhodi ebhalwe umuntu. Ezimbonini ezilawulwayo noma isofthiwe ebalulekile (njengezinhlelo zezokwelapha noma zezindiza), noma yimiphi iminikelo ye-AI ibuyekezwa kanzima.

Alukho uhlelo lwesoftware olujwayelekile namuhla olusetshenziswa ngokuphelele yi-AI kusukela ekuqaleni ngaphandle kokugadwa konjiniyela. Kodwa-ke, okunye ukusetshenziswa okuzimele noma okuzimele kuyavela:

• Ukuhlolwa kweyunithi okuzenzakalelayo: I-AI ingahlaziya ikhodi futhi ikhiqize ukuhlolwa kweyunithi ukuze ihlanganise amacala ahlukahlukene. Uhlaka lokuhlola lungase lukhiqize ngokuzenzakalelayo futhi luqhube lezi zivivinyo ezibhalwe nge-AI ukuze libambe iziphazamisi, lwengeze ukuhlolwa okubhalwe abantu.

• Amapulatifomu anekhodi ephansi/Ayingekho ikhodi ene-AI: Ezinye izinkundla zivumela abangewona abenzi bezinhlelo ukuthi bachaze abakufunayo (isb. “ukwakha ikhasi lewebhu elinefomu lokuxhumana kanye nesizindalwazi ukuze ulondoloze okufakiwe”) futhi isistimu ikhiqiza ikhodi. Ngenkathi isezigabeni zokuqala, lokhu kukhomba esikhathini esizayo lapho i-AI ingazenzela khona isoftware yamacala okusetshenziswa okujwayelekile.

• I-Scripting kanye ne-Glue Code: I-IT automation ngokuvamile ihlanganisa ukubhala izikripthi zokuxhuma izinhlelo. Amathuluzi e-AI ngokuvamile angakhiqiza le mibhalo emincane ngokuzenzakalelayo. Isibonelo, ukubhala umbhalo ukuze uhlaziye ifayela lelogi bese uthumela isaziso se-imeyili - i-AI ingakhiqiza iskripthi esisebenzayo esinokuhlelwa okuncane noma okungekho.

I-Outlook ka-2030-2035: Ibheke “Isoftware Ezithuthukisayo”

Eshumini leminyaka elizayo, i-AI ekhiqizayo kulindeleke ukuthi ithathe ingxenye enkulu yomthwalo wokubhala amakhodi, isondela ekuthuthukisweni kwesoftware ezimele ngokuphelele yezigaba ezithile zamaphrojekthi. Okunye okucatshangelwayo:

• Ukugcwalisa Isici Esiphelele: Ngo-2030, silindele ukuthi i-AI izokwazi ukusebenzisa izici zohlelo lokusebenza ezilula ekugcineni. Umphathi womkhiqizo angase achaze isici ngolimi olulula (“Abasebenzisi kufanele bakwazi ukusetha kabusha iphasiwedi yabo ngesixhumanisi se-imeyili”) futhi i-AI ingase ikhiqize ikhodi edingekayo (ifomu elibheke phambili, i-back-end logic, isibuyekezo sesizindalwazi, ukuthunyelwa kwe-imeyili) futhi iyihlanganise ku-codebase. I-AI ingasebenza ngempumelelo njengonjiniyela omncane ongalandela imininingwane. Unjiniyela ongumuntu angase nje enze ukubuyekezwa kwekhodi futhi enze izivivinyo. Njengoba ukuthembeka kwe-AI kukhula, ukubuyekezwa kwekhodi kungase kube ukushesha okusheshayo uma kunjalo.

• Ukugcinwa Kwekhodi Okuzenzakalelayo: Ingxenye enkulu yobunjiniyela besofthiwe ayigcini nje ngokubhala ikhodi entsha, kodwa ukuvuselela ikhodi ekhona - ukulungisa iziphazamisi, ukuthuthukisa ukusebenza, ukuzivumelanisa nezidingo ezintsha. Abathuthukisi be-AI besikhathi esizayo bazophumelela kulokhu. Uma kunikezwe i-codebase kanye nomyalelo ("uhlelo lwethu lokusebenza luyaphahlazeka uma abasebenzisi abaningi bengena ngasikhathi sinye"), i-AI ingase ithole inkinga (njengesiphazamisi se-concurrency) bese iyichibiyela. Ngo-2035, amasistimu e-AI angaphatha amathikithi okugcina ajwayelekile ngokuzenzakalelayo ngobusuku obubodwa, asebenze njengeqembu elingakhathali lokunakekela amasistimu esoftware.

• Ukuhlanganiswa nokusetshenziswa kwe-API: Njengoba amasistimu esofthiwe engeziwe nama-API eza nemibhalo efundekayo ye-AI, i-ejenti ye-AI ingathola ngokuzimela indlela yokuxhuma i-System A ne-Service B ngokubhala ikhodi yeglue. Isibonelo, uma inkampani ifuna ukuthi uhlelo lwayo lwangaphakathi lwakwa-HR luvunyelaniswe ne-API entsha yokuhola, ingase ibeke i-AI ukuze “yenze lezi zikhulume zodwa,” futhi izobhala ikhodi yokuhlanganisa ngemva kokufunda izici zesistimu zombili.

• Ikhwalithi Nokuthuthukisa: Amamodeli esikhathi esizayo wokukhiqiza ikhodi cishe azohlanganisa ama-loop empendulo ukuze aqinisekise ukuthi ikhodi iyasebenza (isb, sebenzisa izivivinyo noma ukulingisa ku-sandbox). Lokhu kusho ukuthi i-AI ayikwazanga ukubhala ikhodi kuphela kodwa futhi iyazilungisa ngokuyihlola. Ngo-2035, singacabanga nge-AI okuthi, uma inikezwe umsebenzi, iqhubeke iphindaphinda ikhodi kuze kudlule zonke izivivinyo - inqubo umuntu angase angadingi ukuyiqapha umugqa ngomugqa. Lokhu kuzokwandisa kakhulu ukwethenjwa kwekhodi ekhiqizwa ngokuzenzakalelayo.

Umuntu angabona ngeso lengqondo isimo ngo-2035 lapho iphrojekthi yesofthiwe encane - yithi uhlelo lokusebenza lweselula olungokwezifiso lwebhizinisi - lungathuthukiswa kakhulu yi-ejenti ye-AI enikezwe imiyalelo yezinga eliphezulu. “Unjiniyela” ongumuntu kuleso simo ungaphezu komphathi wephrojekthi noma isiqinisekisi, ecacisa izidingo nezingqinamba (ukuphepha, imihlahlandlela yesitayela) nokuvumela i-AI ukuthi iphakamise kanzima ukubhala amakhodi kwangempela.

Kodwa-ke, kuma-software ayinkimbinkimbi, amakhulu (amasistimu okusebenza, ama-algorithms e-AI athuthukile ngokwawo, njll.), ochwepheshe babantu basazobandakanyeka ngokujulile. Idizayini yokudala yokuxazulula izinkinga kanye nesakhiwo sesofthiwe kungenzeka ihlale iholwa umuntu isikhashana. I-AI ingase isingathe imisebenzi eminingi yokubhala amakhodi, kodwa ukunquma ini nokuklama isakhiwo sonke kuyinselele ehlukile. Sekushiwo lokho, njengoba i-AI ekhiqizayo iqala ukusebenzisana - ama-agent amaningi e-AI aphatha izingxenye ezihlukene zesistimu - kungakholakala ukuthi angaklama ngokubambisana izakhiwo ngokwezinga elithile (ngokwesibonelo, i-AI eyodwa iphakamisa idizayini yesistimu, enye iyayigxeka, futhi bayaphindaphinda, nomuntu ongamele inqubo).

Inzuzo enkulu elindelwe ye-AI ekubhaleni amakhodi ukukhulisa umkhiqizo . U-Gartner ubikezela ukuthi ngo-2028, u-90% ngokugcwele wonjiniyela be-software bazobe besebenzisa abasizi bekhodi ye-AI (kusuka ngaphansi kuka-15% ngo-2024) ( GitHub Copilot Tops Research Report on AI Code Assistants -- Visual Studio Magazine ). Lokhu kuphakamisa ukuthi abangaphandle - labo abangasebenzisi i-AI - bazoba mbalwa. Singase futhi sibone ukushoda konjiniyela abangabantu ezindaweni ezithile ezincishiswe i-AI egcwalisa izikhala; empeleni unjiniyela ngamunye angenza okuningi ngomsizi we-AI ongakwazi ukubhala ikhodi ngokuzenzakalelayo.

Ukuthembana kuyohlala kuyindaba esemqoka. Ngisho nango-2035, izinhlangano zizodinga ukuqinisekisa ukuthi ikhodi ekhiqizwe ngokuzenzakalelayo ivikelekile (i-AI akumele yethule ubungozi) futhi iqondaniswe nezinkambiso zomthetho/zezimiso zokuziphatha (isb., i-AI ayifaki ikhodi efihliwe evela kulabhulali yomthombo ovulekile ngaphandle kwelayisensi efanele). Silindele amathuluzi okuphatha e-AI athuthukisiwe angaqinisekisa futhi alandele umsuka wekhodi ebhalwe i-AI ukuze asize ukunika amandla ukubhala ngekhodi okuzimele ngaphandle kobungozi.

Kafushane, maphakathi nawo-2030s, i-AI ekhiqizayo kungenzeka isingethe ingxenye enkulu yokubhala ikhodi yemisebenzi yesofthiwe evamile futhi isize kakhulu kweyinkimbinkimbi. Umjikelezo wokuphila wokuthuthukiswa kwesofthiwe uzozenzakalela kakhulu - kusukela ezidingweni kuye ekusetshenzisweni - nge-AI enamandla okukhiqiza nokukhipha izinguquko zekhodi ngokuzenzakalelayo. Abathuthukisi abangabantu bazogxila kakhulu kumqondo wezinga eliphezulu, ulwazi lomsebenzisi, nokwengamela, kuyilapho ama-agent e-AI egaya ngemininingwane yokusetshenziswa.

I-Generative AI Kusevisi Yekhasimende Nokwesekwa

Uma uke waxhumana nengxoxo yokusekelwa kwamakhasimende eku-inthanethi ezikhathini zakamuva, maningi amathuba okuthi i-AI ibe ngakolunye uhlangothi okungenani ingxenye yayo. Isevisi yamakhasimende iyisizinda esivuthiwe sokuzenzakalela kwe-AI: kuhilela ukuphendula imibuzo yabasebenzisi, iyiphi i-AI ekhiqizayo (ikakhulukazi amamodeli ezingxoxo) engenza kahle kakhulu, futhi ivamise ukulandela imibhalo noma imibhalo eyisisekelo yolwazi, i-AI engayifunda. I-AI ingaphatha kanjani amakhasimende ngokuzimele?

Amakhono Amanje (2025): Ama-Chatbots nama-Virtual Agents Athatha Umugqa Wangaphambili

Kusukela namuhla, izinhlangano eziningi zisebenzisa i-AI chatbots njengendawo yokuqala yokuxhumana enkonzweni yamakhasimende. Lokhu kusuka kuma-bots asekelwe emthethweni (“Cindezela 1 ukuze uthole inkokhiso, 2 ukuze uthole ukwesekwa…”) kuya kuma-chatbot akhiqizayo e-AI athuthukisiwe angatolika imibuzo yamahhala futhi aphendule ngengxoxo. Amaphuzu abalulekile:

• Ukusingatha Imibuzo Ejwayelekile: Abenzeli be-AI benza kahle kakhulu ekuphenduleni imibuzo evame ukubuzwa, banikeze ulwazi (amahora esitolo, izinqubomgomo zokubuyiselwa kwemali, izinyathelo zokuxazulula izinkinga zezinkinga ezaziwayo), kanye nokuqondisa abasebenzisi ngezinqubo ezijwayelekile. Isibonelo, i-chatbot ye-AI yebhange ingasiza ngokuzenzakalelayo umsebenzisi ukuthi ahlole ibhalansi ye-akhawunti yakhe, asethe kabusha iphasiwedi, noma achaze indlela yokufaka isicelo semali mboleko, ngaphandle kosizo lomuntu.

• Ukuqonda Ulimi Lwemvelo: Amamodeli akhiqizayo esimanje avumela uketshezi oluthe xaxa kanye nokusebenzisana "okufana nomuntu". Amakhasimende angabhala umbuzo ngamagama awo futhi i-AI ingakwazi ukubamba inhloso. Izinkampani zibika ukuthi ama-ejenti e-AI anamuhla anelisa kakhulu kumakhasimende kunama-clunky bots eminyaka embalwa edlule - cishe isigamu samakhasimende manje akholelwa ukuthi ama-ejenti e-AI angaba nozwela futhi aphumelele lapho ebhekana nokukhathazeka ( 59 AI izibalo zesevisi yamakhasimende zika-2025 ), okubonisa ukwethenjwa okukhulayo kusevisi eqhutshwa yi-AI.

• Ukusekelwa Kweziteshi Eziningi: I-AI ayikho nje engxoxweni. Abasizi bezwi (njengezinhlelo ze-IVR zocingo ezine-AI ngemuva kwabo) sebeqala ukuphatha izingcingo, futhi i-AI ingaphinda ibhale izimpendulo ze-imeyili kumibuzo yamakhasimende engase iphume ngokuzenzakalelayo uma ibonakala inembile.

• Lapho Abantu Bengena: Imvamisa, uma i-AI idideka noma umbuzo uyinkimbinkimbi kakhulu, izodluliselwa kumenzeli ongumuntu. Izinhlelo zamanje zinhle ekwazini imikhawulo yazo ezimweni eziningi. Ngokwesibonelo, uma ikhasimende libuza okuthile okungajwayelekile noma libonisa ukukhungatheka (“Lesi yisikhathi sesithathu ngixhumana nawe futhi ngiphatheke kabi…”), i-AI ingase ihlabe umkhosi lokhu ukuze kuthathwe umuntu. Umkhawulo wokunikezela ubekwe yizinkampani ukuze zilinganisele ukusebenza kahle nokwaneliseka kwamakhasimende.

Izinkampani eziningi zibike izingxenye ezibalulekile zokusebenzisana ezixazululwa yi-AI kuphela. Ngokwenhlolovo yezimboni, cishe i-70-80% yemibuzo yamakhasimende evamile ingasingathwa yi-AI chatbots namuhla, futhi cishe u-40% wokusebenzelana kwamakhasimende ezinkampani kuzo zonke iziteshi sekuvele kuzenzakalelayo noma kusizwa nge-AI ( 52 AI Customer Service Statistics Okufanele Uzazi - Plivo ). I-Global AI Adoption Index ye-IBM (2022) ikhombise u-80% wezinkampani ukuthi zisebenzisa noma zihlele ukusebenzisa ama-chatbots e-AI ngenkonzo yamakhasimende ngo-2025.

Intuthuko ethokozisayo ukuthi i-AI ayigcini nje ngokuphendula amakhasimende, kodwa isiza ngokuqhubekayo ama-agent angabantu ngesikhathi sangempela. Isibonelo, phakathi nengxoxo ebukhoma noma ikholi, i-AI ingase ilalele futhi inikeze umenzeli ongumuntu izimpendulo eziphakanyisiwe noma ulwazi olufanele ngokushesha. Lokhu kufiphaza umugqa wokuzimela - i-AI ayibheki ikhasimende lodwa, kodwa ibandakanyeka ngokukhuthele ngaphandle kombuzo ocacile womuntu. Isebenza ngempumelelo njengomeluleki ozimele we-ejenti.

I-Outlook ka-2030-2035: Ukusebenzelana Kwamakhasimende Okuqhutshwa yi-AI

Ngo-2030, iningi lokusebenzelana kwamakhasimende kulindeleke ukuthi lifake i-AI, kanti okuningi kuzosingathwa yi-AI kusukela ekuqaleni kuya ekugcineni. Izibikezelo namathrendi asekela lokhu:

• Imibuzo Ephakeme Eyinkimbinkimbi Ixazululiwe: Njengoba amamodeli e-AI ehlanganisa ulwazi oluningi futhi athuthukisa ukucabanga, azokwazi ukusingatha izicelo zamakhasimende eziyinkimbinkimbi. Esikhundleni sokuphendula nje "Ngiyibuyisela kanjani into?", I-AI yesikhathi esizayo ingase isingathe izinkinga zezinyathelo eziningi ezifana nokuthi, "I-inthanethi yami iphansi, ngizamile ukuqalisa kabusha, ungasiza?" ngokuxilonga inkinga ngengxoxo, ukuqondisa ikhasimende ekuxazululeni izinkinga ezithuthukile, futhi kuphela uma konke okunye kwehluleka ukuhlela uchwepheshe - imisebenzi namuhla engadinga ubuchwepheshe bosekelo lomuntu. Enkonzweni yamakhasimende okunakekelwa kwezempilo, i-AI ingase isingathe ukuhlela isikhathi sokuqokwa kwesiguli noma imibuzo yomshwalense ekugcineni.

• Isixazululo Sesevisi Yokuphela-kuya-Ekupheleni: Singabona i-AI ingagcini nje ngokutshela ikhasimende ukuthi lenzeni, kodwa empeleni siyenzela ikhasimende ngaphakathi kwezinhlelo ezingemuva. Isibonelo, uma ikhasimende lithi “Ngifuna ukushintsha indiza yami ibe uMsombuluko olandelayo bese ngengeza esinye isikhwama,” umenzeli we-AI ngo-2030 angase axhumane ngokuqondile nesistimu yokubhukha yenkampani yezindiza, enze ushintsho, acubungule inkokhelo yesikhwama, futhi aqinisekise ekhasimendeni – konke ngokuzenzakalelayo. I-AI iba i-ejenti yesevisi egcwele, hhayi nje umthombo wolwazi.

• Ama-Omnipresent AI Agents: Izinkampani cishe zizosebenzisa i-AI kuzo zonke izindawo zokuthinta amakhasimende - ifoni, ingxoxo, i-imeyili, inkundla yezokuxhumana. Amakhasimende amaningi angase angaboni nokuthi akhuluma ne-AI noma umuntu, ikakhulukazi njengoba amazwi e-AI eba ngokwemvelo futhi ingxoxo iphendula ngokuqaphela umongo. Ngo-2035, ukuxhumana nesevisi yamakhasimende ngokuvamile kungasho ukusebenzisana ne-AI ehlakaniphile ekhumbula ukusebenzisana kwakho kwesikhathi esidlule, eqondayo okuncamelayo, futhi evumelanisa nethoni yakho - empeleni i-ejenti yomuntu siqu yawo wonke amakhasimende.

• Ukuthatha Izinqumo kwe-AI Ekusebenzisaneni: Ngaphandle kokuphendula imibuzo, i-AI izoqala ukwenza izinqumo ezidinga ukugunyazwa kwabaphathi njengamanje. Ngokwesibonelo, namuhla umenzeli ongumuntu angase adinge imvume yomphathi ukuze anikeze imbuyiselo noma isaphulelo esikhethekile ukuze ashweleze ikhasimende elithukuthele. Ngokuzayo, i-AI ingase iphathiswe lezo zinqumo, ngaphakathi kwemikhawulo ebekiwe, ngokusekelwe enanini lempilo yekhasimende elibaliwe kanye nokuhlaziya imizwelo. Ucwaningo olwenziwa yi-Futurum/IBM lwabikezela ukuthi ngo-2030 cishe u-69% wezinqumo ezenziwa ngesikhathi sangempela sokuxoxisana namakhasimende zizokwenziwa imishini ehlakaniphile ( Ukuze Ucabange Kabusha I-Shift ukuya ku-CX, Abakhangisi Kumele Benze Lezi Zinto Ezi-2 ) - ngokuphumelelayo i-AI inquma inkambo engcono kakhulu yesenzo ekuxhumaneni.

• Ukubandakanyeka kwe-AI okungu-100%: Umbiko owodwa uphakamisa ukuthi i-AI izogcina ibambe iqhaza kukho konke ukusebenzisana kwamakhasimende ( izibalo zesevisi yamakhasimende ezingama-59 AI zika-2025 ), kungakhathaliseki ukuthi zingaphambili noma ngemuva. Lokho kungase kusho ukuthi ngisho noma umuntu esebenzisana nekhasimende, bazosizwa i-AI (enikeza iziphakamiso, ukubuyisa ulwazi). Kungenjalo, incazelo iwukuthi awukho umbuzo wekhasimende ongaphendulwa nganoma yisiphi isikhathi - uma abantu bengaxhunyiwe ku-inthanethi, i-AI ihlala ikhona.

Ngo-2035, singase sithole ama-ejenti esevisi yamakhasimende abantu asebekhethekile kuphela ngezimo ezibucayi kakhulu noma ezithinta inhliziyo (isb., amaklayenti e-VIP noma ukuxazululwa kwezikhalazo eziyinkimbinkimbi ezidinga uzwela lomuntu). Imibuzo ejwayelekile - ukusuka ebhange kuye ekuthengiseni kuye ekusekelweni kwezobuchwepheshe - ingasizwa uxhaxha lwama-ejenti e-AI asebenza 24/7, efunda ngokuqhubekayo ekuhlanganyeleni ngakunye. Lolu shintsho lungenza isevisi yamakhasimende ingaguquguquki futhi isheshe, njengoba i-AI ingabagcini abantu belindile futhi ingakwazi ukwenza imisebenzi eminingi ukuze isingathe amakhasimende angenamkhawulo ngasikhathi sinye.

Kunezinselelo okufanele uzinqobe kulo mbono: I-AI kufanele iqine kakhulu ukuze isingathe ukungaqiniseki kwamakhasimende angabantu. Kumelwe ikwazi ukubhekana nenkulumo-ze, intukuthelo, ukudideka, nezinhlobonhlobo zezindlela abantu abakhuluma ngazo. Idinga nolwazi lwakamuva (akukho phuzu uma imininingwane ye-AI isiphelelwe yisikhathi). Ngokutshala imali ekuhlanganisweni phakathi kwe-AI nesizindalwazi senkampani (ngolwazi lwesikhathi sangempela kuma-oda, ukuphuma, njll.), lezi zithiyo zingaxazululwa.

Ngokufanelekile, izinkampani kuzodingeka zinqume ukuthi zizodalula nini ukuthi “ukhuluma ne-AI” futhi ziqinisekise ubulungisa (i-AI ayiphathi amakhasimende athile ngendlela ehlukile ngendlela engafanele ngenxa yokuqeqeshwa okuchemile). Uma sicabanga ukuthi lezi ziphethwe, icala lebhizinisi liqinile: Isevisi yamakhasimende ye-AI inganciphisa kakhulu izindleko kanye nezikhathi zokulinda. Imakethe ye-AI ekuhlinzekeni amakhasimende kulindeleke ukuthi ikhule ifinyelele emashumini ezigidigidi zamadola ngo-2030 ( I-AI Embikweni Wemakethe Yesevisi Yabathengi 2025-2030: Ikesi ) ( Indlela I-Generative AI Ithuthukisa Ngayo Logistics | Ryder ) njengoba izinhlangano zitshala imali kulawa makhono.

Kafushane, lindela ikusasa lapho insizakalo yamakhasimende e-AI ezimele iyinto evamile . Ukuthola usizo ngokuvamile kuzosho ukusebenzisana nomshini ohlakaniphile ongaxazulula inkinga yakho ngokushesha. Abantu basazoba semgwaqweni wokwengamela nokuphatha amacala asemaphethelweni, kodwa kakhulu njengabaphathi babasebenzi be-AI. Umphumela ungase ube isevisi esheshayo, eyenzelwe umuntu siqu kakhudlwana kubathengi - inqobo nje uma i-AI iqeqeshwe kahle futhi igadiwe ukuze kuvinjelwe ukukhungatheka kokuhlangenwe nakho "kwe-robot hotline" yesikhathi esidlule.

I-Generative AI ku-Healthcare and Medicine

Ukunakekelwa kwezempilo kuyinkambu lapho izingozi ziphezulu. Umqondo we-AI osebenza ngaphandle kokuqondisa komuntu kwezokwelapha uvusa kokubili injabulo (yokusebenza kahle nokufinyelela) kanye nokuqapha (ngenxa yezizathu zokuphepha nozwela). I-Generative AI isiqalile ukungena ezindaweni ezinjengokuhlaziywa kwezithombe zezokwelapha, imibhalo yomtholampilo, kanye nokutholwa kwezidakamizwa. Yini engayenza ngokuzibophezela ngokwayo?

Amakhono Amanje (2025): Ukusiza Odokotela, Hhayi Ukuwashintsha

Njengamanje, i-AI ekhiqizayo ekunakekelweni kwezempilo isebenza njengomsizi onamandla kochwepheshe bezokwelapha, esikhundleni sokuba ngumenzi wezinqumo ozimele. Ngokwesibonelo:

• Imibhalo Yezokwelapha: Okunye okuphumelela kakhulu ukuthunyelwa kwe-AI ekunakekelweni kwezempilo ukusiza odokotela ngamaphepha. Amamodeli olimi lwemvelo angabhala ukuvakasha kwesiguli futhi akhiqize amanothi omtholampilo noma izifinyezo zokukhishwa. Izinkampani “zinababhali be-AI” abalalela ngesikhathi sokuhlolwa (ngemakrofoni) futhi bakhiqize ngokuzenzakalela uhlaka lwamanothi okuhlangana ukuze udokotela alubuyekeze. Lokhu kongela odokotela isikhathi sokuthayipha. Amanye amasistimu aze agcwalise ngokuzenzakalela izingxenye zamarekhodi ezempilo e-elekthronikhi. Lokhu kungenziwa ngokungenelela okuncane - udokotela uvele alungise noma yimaphi amaphutha amancane kuhlaka, okusho ukuthi ukubhala inothi kuzimele.

• I-Radiology ne-Imaging: I-AI, okuhlanganisa amamodeli akhiqizayo, ingahlaziya ama-X-reyi, ama-MRIs, nezikena ze-CT ukuze ithole okudidayo (njengamathumba noma ukuphuka). Ngo-2018, i-FDA yagunyaza uhlelo lwe-AI lokutholwa okuzimele kwe-retinopathy yesifo sikashukela (isimo samehlo) ezithombeni ze-retina - ikakhulukazi, yayigunyazwe ukwenza ikholi ngaphandle kokubuyekezwa kochwepheshe kulowo mongo othile wokuhlola. Leyo sistimu bekungeyona i-AI yokukhiqiza, kodwa ibonisa ukuthi abalawuli bavumele ukuxilongwa kwe-AI okuzenzakalelayo ezimeni ezilinganiselwe. Amamodeli akhiqizayo ayasebenza ukuze kwakheke imibiko ephelele. Isibonelo, i-AI ingase ihlole i-X-ray yesifuba bese ibhala umbiko we-radiologist othi “Akukho okutholakele okubucayi. Amaphaphu acacile. Inhliziyo isayizi evamile.” Udokotela we-radiologist ube eseqinisekisa futhi asayine. Kwezinye izimo ezijwayelekile, le mibiko ingase iphume ngaphandle kokuhlelwa uma isazi se-radiologist sithemba i-AI futhi sivele sihlole ngokushesha.

• Izihloli Zezimpawu Nabahlengikazi Ababonakalayo: Ama-chatbots e-AI akhiqizayo asetshenziswa njengezihloli zezimpawu eziphambili. Iziguli zingafaka izimpawu zazo futhi zithole iseluleko (isb, “Kungase kube umkhuhlane ovamile; ukuphumula nokuphuza, kodwa bonana nodokotela uma kwenzeka u-X noma u-Y.”). Izinhlelo zokusebenza ezifana ne-Babylon Health zisebenzisa i-AI ukunikeza izincomo. Okwamanje, lezi zivame ukufakwa kuhlaka njengeseluleko sezokwelapha esinolwazi, hhayi esiqondile, futhi zikhuthaza ukulandelwa nodokotela wesintu ngezinkinga ezibucayi.

• Ukutholwa Kwezidakamizwa (I-Generative Chemistry): Amamodeli e-Generative AI angaphakamisa ukwakheka kwamangqamuzana amasha ezidakamizwa. Lokhu kuningi esizindeni socwaningo kunokunakekelwa kwesiguli. Lawa ma-AI asebenza ngokuzenzakalelayo ukuphakamisa izinkulungwane zamakhompiyutha abantu abazohlangana nezinto ezifiselekayo, osokhemisi abangabantu bese bezibuyekeza futhi bazihlole elebhu. Izinkampani ezifana ne-Insilico Medicine zisebenzise i-AI ukukhiqiza abantu abasha bezidakamizwa ngesikhathi esincane kakhulu. Nakuba lokhu kungasebenzisani ngokuqondile neziguli, kuyisibonelo se-AI eyenza ngokuzenzakalelayo izixazululo (imiklamo ye-molecule) abantu abebezothatha isikhathi eside ukuyithola.

• Imisebenzi Yokunakekelwa Kwezempilo: I-AI isiza ekulungiseleleni ukuhlela, ukuphathwa kokuhlinzeka, nokunye ukuhanjiswa ezibhedlela. Isibonelo, imodeli ekhiqizayo ingase ilingise ukugeleza kwesiguli futhi iphakamise ukuhlela ukulungiswa ukuze kuncishiswe izikhathi zokulinda. Yize zingabonakali kangako, lezi yizinqumo i-AI engazenza ngezinguquko ezincane ezenziwa ngesandla.

Kubalulekile ukusho ukuthi kusukela ngo-2025, asikho isibhedlela esivumela i-AI ngokuzimela ukuthi yenze izinqumo ezinkulu zezokwelapha noma ukwelashwa ngaphandle kokusayina komuntu. Ukuxilongwa nokuhlelwa kokwelashwa kuhlala kuqinile ezandleni zomuntu, i-AI ihlinzeka ngokufaka. Ukwethenjwa okudingekayo ukuze i-AI itshele isiguli ngokugcwele ukuthi “Unomdlavuza” noma ukusinikeza imithi akukabi khona, futhi akufanele kube ngaphandle kokuqinisekisa okubanzi. Ochwepheshe bezokwelapha basebenzisa i-AI njengepheya lesibili lamehlo noma njengethuluzi lokonga isikhathi, kodwa baqinisekisa imiphumela ebucayi.

I-Outlook ka-2030-2035: I-AI Njengozakwethu Kadokotela (futhi mhlawumbe Umhlengikazi noma Usokhemisi)

Eshumini leminyaka elizayo, silindele ukuthi i-AI ekhiqizayo ithathe imisebenzi yomtholampilo evamile ngokuzimela futhi ithuthukise ukufinyeleleka kwezinsizakalo zokunakekelwa kwezempilo:

• Ukuxilongwa Kokuqala Okuzenzakalelayo: Ngo-2030, i-AI yayingakwazi ukuphatha ngokuthembekile ukuhlaziywa kokuqala kwezimo eziningi ezivamile. Cabanga ngohlelo lwe-AI emtholampilo olufunda izimpawu zesiguli, umlando wezokwelapha, ngisho nethoni yaso nezimpawu zobuso ngekhamera, futhi lunikeze isiphakamiso sokuxilonga nokuhlola okunconyiwe - konke ngaphambi kokuba udokotela ongumuntu aze abone isiguli. Udokotela angabe esegxila ekuqinisekiseni nasekuxoxeni ngokuxilongwa. Ku-telemedicine, isiguli singase siqale sixoxe ne-AI enciphisa inkinga (isb., ukutheleleka okungenzeka kwe-sinus ngokumelene nento enzima kakhulu) bese ibaxhumanisa nodokotela uma kudingeka. Abalawuli bangase bavumele i-AI ukuthi ngokusemthethweni izimo ezithile ezincane ngaphandle kokuqondisa komuntu uma kufakazelwa ukuthi zinembile ngokwedlulele - isibonelo, i-AI ehlonza ukutheleleka kwendlebe okuqondile okuvela esithombeni se-otoscope kungenzeka.

• Iziqapheli Zezempilo Zomuntu Siqu: Ngokwanda kwezinto ezigqokwayo (amawashi ahlakaniphile, izinzwa zezempilo), i-AI izoqapha iziguli ngokuqhubekayo futhi ngokuzimele ixwayise ngezinkinga. Isibonelo, ngo-2035 i-AI yakho egqokekayo ingase ibone isigqi senhliziyo esingavamile futhi ikuhlelele ngokuzenzakalelayo ukubonana okuphuthumayo noma ishayele i-ambulensi uma ithola izimpawu zokuhlaselwa yinhliziyo noma unhlangothi. Lokhu kudlulela endaweni yesinqumo esizimele - kunquma ukuthi isimo siyisimo esiphuthumayo futhi siyasebenza - okungaba ukusetshenziswa kwe-AI okungenzeka futhi kusindise ukuphila.

• Izincomo Zokwelashwa: I-Generative AI eqeqeshwe ngezincwadi zezokwelapha kanye nedatha yesiguli ingase iphakamise izinhlelo zokwelashwa eziqondene nawe. Ngo-2030, ezifweni eziyinkimbinkimbi njengomdlavuza, amabhodi wesimila we-AI ayengakwazi ukuhlaziya ukwakheka kofuzo lwesiguli nomlando wezokwelapha futhi abhale ngokuzimele uhlobo lokwelapha olunconyiwe (uhlelo lwe-chemo, ukukhethwa kwezidakamizwa). Odokotela abangabantu bazoyibuyekeza, kodwa ngokuhamba kwesikhathi njengoba ukuzethemba kukhula, bangase baqale ukwamukela izinhlelo ezikhiqizwe yi-AI ikakhulukazi ezimweni ezivamile, balungise kuphela lapho kudingekile.

• Abahlengikazi Ababonakalayo Nokunakekelwa Kwasekhaya: I-AI engaxoxa futhi inikeze isiqondiso sezokwelapha ingase isingathe ukulandelela okuningi nokuqapha kokunakekelwa okungapheli. Isibonelo, iziguli ekhaya ezinezifo ezingelapheki zingabika amamethrikhi nsuku zonke kumsizi womhlengikazi we-AI onikeza iseluleko (“Ushukela osegazini lakho uphezulu kancane, cabanga ukulungisa ukudla kwakho kwakusihlwa”) kanye nokungena kuphela kumhlengikazi ongumuntu lapho ukufundwa kungaphandle kwebanga noma kuphakama izinkinga. Le AI ingasebenza ngokuzimele ngaphansi kokugadwa okukude kukadokotela.

• I-Medical Imaging and Lab Analysis – Amapayipi Asebenza Ngokugcwele: Ngo-2035, ukufunda izikena zezokwelapha kungase kwenziwe kakhulukazi i-AI kweminye imikhakha. Izazi zeRadiologists zizokwengamela izinhlelo ze-AI futhi zibambe amacala ayinkimbinkimbi, kodwa iningi lezikena ezijwayelekile (ezijwayelekile ngempela) "zingafundwa" futhi zisayinwe yi-AI ngokuqondile. Ngokufanayo, ukuhlaziya amaslayidi e-pathology (isibonelo, ukuthola amaseli omdlavuza ku-biopsy) kungenziwa ngokuzimele ukuze kuhlolwe okokuqala, kusheshise imiphumela yelebhu.

• Ukutholwa Kwezidakamizwa kanye Nezilingo Zomtholampilo: I-AI cishe ngeke idizayine kuphela ama-molecule ezidakamizwa kodwa futhi ikhiqize idatha yokwenziwa yesiguli yezivivinyo noma ithole amakhandidethi wokuhlola afanele. Ingase iqhube ngokuzenzakalelayo izivivinyo ezibonakalayo (ukulingisa ukuthi iziguli zizosabela kanjani) ukuze unciphise izinketho ngaphambi kokuhlolwa kwangempela. Lokhu kungaletha imithi ukuthi imakethwe ngokushesha ngokuhlolwa okuncane okuqhutshwa abantu.

Umbono kadokotela we-AI othatha isikhundla sikadokotela ongumuntu usekude futhi useyimpikiswano. Ngisho nango-2035, okulindelekile ukuthi i-AI izosebenza njengozakwethu kodokotela esikhundleni sokuthatha indawo yokuthintwa komuntu. Ukuxilongwa okuyinkimbinkimbi kuvame ukudinga ukuqondisisa, izimiso zokuziphatha, nezingxoxo ukuze uqonde umongo wesiguli - izindawo lapho odokotela babantu bephumelela khona. Sekushiwo lokho, i-AI ingase isingathe, ithi, i-80% yomsebenzi ovamile: amaphepha, amacala aqondile, ukuqapha, njll., okuvumela odokotela abangabantu ukuthi bagxile ku-20% okhohlisayo nasebudlelwaneni besiguli.

Kunezithiyo ezibalulekile: ukugunyazwa kokulawula kwe-AI ezimele ekunakekelweni kwezempilo kuqinile (ngokufanelekile). Amasistimu e-AI azodinga ukuqinisekiswa okubanzi komtholampilo. Singase sibone ukwamukelwa okwandayo - isb., i-AI ivunyelwe ukuxilonga noma ukwelapha ngokuzenzakalelayo ezindaweni lapho kungekho dokotela abatholakala khona, njengendlela yokwelula ukufinyelela kwezokunakekelwa kwezempilo (cabanga “ngomtholampilo we-AI” osedolobhaneni elikude ngo-2030 osebenza ngokugadwa ngocingo ngezikhathi ezithile okuvela kudokotela edolobheni).

Ukucatshangelwa kwezimiso zokuziphatha kubukeka kukhulu. Ukuziphendulela (uma i-AI ezimele yenza iphutha ekuxilongweni, ubani onesibopho?), imvume enolwazi (iziguli zidinga ukwazi ukuthi i-AI iyabandakanyeka yini ekunakekelweni kwazo), kanye nokuqinisekisa ukulingana (i-AI isebenza kahle kuzo zonke izigaba zabantu, ukugwema ukuchema) kuyizinselele zokuzulazula. Uma sicabanga ukuthi lezo sezixazululiwe, maphakathi nawo-2030 i-AI ekhiqizayo ingase ihlanganiswe nendwangu yokulethwa kokunakekelwa kwezempilo, yenze imisebenzi eminingi ekhulula abahlinzeki babantu futhi okungenzeka ifinyelele iziguli ezinokufinyelela okulinganiselwe okwamanje.

Kafushane, ngo-2035 ukunakekelwa kwezempilo cishe kuzobona i-AI ihlanganiswe ngokujulile kodwa ikakhulukazi ngaphansi kwe-hood noma ezindimeni ezisekelayo. Sizothemba i-AI ukuthi izokwenza okuningi iyodwa - funda izikena, ezibalulekile zewashi, izinhlelo ezisalungiswa - kodwa kunenetha yokuphepha yokuqondisa komuntu okusasebenza ukuze kuthathwe izinqumo ezibalulekile. Umphumela ungaba uhlelo lokunakekelwa kwezempilo olusebenza kahle kakhulu, olusabelayo, lapho i-AI iphatha khona ukuphakamisa okusindayo futhi abantu banikeze uzwela nesahlulelo sokugcina.

I-Generative AI Kwezemfundo

Imfundo ingenye inkambu lapho i-AI ekhiqizayo yenza amagagasi, kusukela ku-AI-powered tutoring bots kuya ekugredini okuzenzakalelayo nokudalwa kokuqukethwe. Ukufundisa nokufunda kubandakanya ukuxhumana nokudala, okungamandla amamodeli okukhiqiza. Kodwa ingabe i-AI ingethenjwa ukuthi izofundisa ngaphandle kokugadwa uthisha?

Amakhono Amanje (2025): Abafundisi kanye Nabakhiqizi Bokuqukethwe ku-Leash

Njengamanje, i-AI isetshenziswa kwezemfundo ngokuyinhloko njengethuluzi lokwengeza esikhundleni sokuba uthisha ozimele. Izibonelo zokusetshenziswa kwamanje:

• I-AI Tutoring Assistants: Amathuluzi afana ne-“Khanmigo” ye-Khan Academy (enikwa amandla yi-GPT-4) noma izinhlelo zokusebenza ezihlukahlukene zokufunda ulimi zisebenzisa i-AI ukuze zilingise umfundisi oyedwa noma oyedwa noma uzakwethu wengxoxo. Abafundi bangabuza imibuzo ngolimi lwemvelo futhi bathole izimpendulo noma izincazelo. I-AI inganikeza izeluleko zezinkinga zomsebenzi wasekhaya, ichaze imiqondo ngezindlela ezihlukene, noma idlale indima njengomuntu ongokomlando wesifundo somlando osebenzisanayo. Kodwa-ke, laba bafundisi be-AI bavame ukusetshenziswa ngokwengamela; othisha noma abagcini bohlelo lokusebenza bavame ukuqapha izingxoxo noma babeke imingcele kulokho i-AI engaxoxa ngakho (ukugwema ulwazi olunganembile noma okuqukethwe okungalungile).

• Ukudalwa Kokuqukethwe Kothisha: I-Generative AI isiza othisha ngokudala imibuzo yemibuzo, izifinyezo zokufunda, uhlaka lwezinhlelo zezifundo, nokunye. Uthisha angase abuze i-AI, “Khiqiza izinkinga zokuprakthiza ezi-5 kuma-quadratic equations ngezimpendulo,” konga isikhathi sokulungiselela. Lokhu ukukhiqizwa kokuqukethwe okuzimele, kodwa uthisha ngokuvamile ubuyekeza okukhiphayo ukuze athole ukunemba nokuqondanisa nekharikhulamu. Ngakho-ke ingaphezu kwedivayisi yokonga abasebenzi kunokuzimela ngokugcwele.

• Ukugreda Nempendulo: I-AI ingakwazi ukukala ngokuzenzakalelayo izivivinyo zokukhetha eziningi (akukho okusha lapho) futhi ngokwandayo ingahlola izimpendulo ezimfushane noma izindatshana. Amanye amasistimu esikole asebenzisa i-AI ukuze alinganisele izimpendulo ezibhaliwe futhi anikeze impendulo kubafundi (isb., ukulungisa uhlelo lolimi, iziphakamiso zokwandisa impikiswano). Nakuba kungewona umsebenzi okhiqizayo, ama-AI amasha angakwazi nokwenza umbiko womuntu siqu wempendulo yomfundi ngokusekelwe ekusebenzeni kwakhe, egqamisa izindawo okufanele zithuthukiswe. Othisha bavamise ukuhlola kabili izindatshana ezisezingeni le-AI kulesi sigaba ngenxa yokukhathazeka ngama-nuance.

• Amasistimu Wokufunda Aguquguqukayo: Lezi izinkundla ezilungisa ubunzima noma isitayela sezinto ezibonakalayo ngokusekelwe ekusebenzeni komfundi. I-Generative AI ithuthukisa lokhu ngokudala izinkinga ezintsha noma izibonelo kumpukane eziklanyelwe izidingo zomfundi. Isibonelo, uma umfundi edonsa kanzima ngombono, i-AI ingase ikhiqize esinye isifaniso noma umbuzo wokuzijwayeza ogxile kulowo mqondo. Lokhu kuzimele ngandlela thize, kodwa kungaphakathi kwesistimu eklanywe othisha.

• Ukusetshenziswa Komfundi Ukufunda: Abafundi ngokwabo basebenzisa amathuluzi afana ne-ChatGPT ukuze basize ekufundeni - bacele ukucaciswa, ukuhumusha, noma ngisho nokusebenzisa i-AI ukuze bathole impendulo ngohlaka lwendatshana (“thuthukisa isigaba sami sesingeniso”). Lokhu kuziqondisa futhi kungaba ngaphandle kolwazi lukathisha. I-AI kulesi simo isebenza njengomfundisi odingeka kakhulu noma isifundi sokulungisa iphutha. Inselele iwukuqinisekisa ukuthi abafundi bayisebenzisela ukufunda kunokuba bathole izimpendulo kuphela (ubuqotho bezemfundo).

Kuyacaca ukuthi kusukela ngo-2025, i-AI kwezemfundo inamandla kodwa ijwayele ukusebenza nothisha ongumuntu ku-loop ohlola iminikelo ye-AI. Kunesexwayiso esiqondakalayo: asifuni ukuthemba i-AI ukuthi izofundisa ulwazi olungalungile noma iphathe ukuxhumana kwabafundi okubucayi endaweni engenalutho. Othisha babheka abafundisi be-AI njengabasizi abawusizo abanganikeza abafundi ukuzilolonga okwengeziwe nezimpendulo ezisheshayo zemibuzo evamile, bakhulule othisha ukuthi bagxile ekufundiseni okujulile.

I-Outlook ka-2030-2035: Abafundisi Be-AI Abazenzela Umuntu Nezinsiza Zokufundisa Ezizenzakalelayo

Eshumini leminyaka elizayo, silindele ukuthi i-AI ekhiqizayo izovumela ukuzizwisa komuntu siqu nokuzenzela okuzimele , kuyilapho izindima zothisha zishintsha:

• I-AI Personal Tutors Yawo Wonke Umfundi: Ngo-2030, umbono (owabelanwe ngochwepheshe abafana no-Sal Khan we-Khan Academy) owokuthi umfundi ngamunye angaba nokufinyelela kumfundisi we-AI osebenza ngempumelelo njengomfundisi ongumuntu ezicini eziningi ( Lo mfundisi we-AI angenza abantu bahlakaniphe izikhathi eziyi-10, umdali wakhe uthi ). Laba bafundisi be-AI bazotholakala 24/7, bazi umlando wokufunda womfundi ngokuseduze, futhi baguqule isitayela sabo sokufundisa ngokufanele. Isibonelo, uma umfundi engumfundi obonakalayo onenkinga yomqondo we-algebra, i-AI ingase idale incazelo ebonakalayo noma ukulingisa okusebenzisanayo ukusiza. Ngenxa yokuthi i-AI ingakwazi ukulandelela inqubekelaphambili yomfundi ngokuhamba kwesikhathi, ingakwazi ngokuzenzakalelayo ukunquma ukuthi yisiphi isihloko okufanele sibuyekezwe ngokulandelayo noma ukuthi uthuthukela nini ekhonweni elisha – iphathe ngempumelelo uhlelo lwesifundo lwalowo mfundi ngomqondo omncane.

• Umthwalo Owehlisiwe Womsebenzi Wothisha EmiSebenzini Ejwayelekile: Ukuhlunga, ukwenza amaphepha okusebenzela, ukubhala izinto zokufunda - le misebenzi ingase icishe ikhishwe ngokuphelele ku-AI ngo-2030. I-AI ingase ikhiqizele ikilasi umsebenzi wasekhaya ofanele isonto lonke, ibale yonke imisebenzi yesonto eledlule (ngisho nevaliwe) ngempendulo, futhi iqokomise kuthisha ukuthi yiziphi abafundi abangase badinge usizo olwengeziwe mayelana nokuthi yiziphi izihloko. Lokhu kungenzeka ngokufaka okuncane kothisha, mhlawumbe ukubuka nje ngokushesha ukuze uqinisekise ukuthi amamaki e-AI abonakala efanelekile.

• Ama-Autonomous Adaptive Learning Platforms: Singase sibone ngokugcwele izifundo eziqhutshwa yi-AI zezifundo ezithile. Cabanga ngesifundo esiku-inthanethi esingenaye umfundisi ongumuntu lapho i-ejenti ye-AI yethula izinto ezibonakalayo, inikeze izibonelo, iphendule imibuzo, futhi ilungise ijubane ngokusekelwe kumfundi. Okuhlangenwe nakho komfundi kungase kuhluke kubo, kwenziwe ngesikhathi sangempela. Okunye ukuqeqeshwa kwezinkampani nokufunda kwabantu abadala kungase kudlulele kulo modeli ngokushesha, lapho ngo-2035 isisebenzi singase sithi “Ngifuna ukufunda ama-Excel macros athuthukisiwe” futhi umfundisi we-AI ezowafundisa ngohlelo lwezifundo oluqondene nomuntu siqu, okuhlanganisa ukuvivinya umzimba kokukhiqiza nokuhlola izixazululo zabo, ngaphandle komqeqeshi ongumuntu.

• Abasizi Bekilasi Le-AI: Emakilasini aphathekayo noma abonakalayo, i-AI ingalalela izingxoxo zekilasi futhi isize uthisha endiza (isb, iziphakamiso zokuhleba ngomsindo ozwayo: “Abafundi abambalwa babukeka bedidekile ngalowo mqondo, mhlawumbe banikeze esinye isibonelo”). Ingase futhi yengamele izinkundla zamakilasi aku-inthanethi, iphendule imibuzo eqondile ebuzwa ngabafundi (“Ufanele umsebenzi ozokwenziwa nini?” noma icacise iphuzu lokufundisa) ukuze uthisha angahlatshwa ama-imeyili. Ngo-2035, ukuba nothisha we-AI ekamelweni, kuyilapho uthisha ongumuntu egxile ekuqondisweni kwezinga eliphezulu nasezicini ezikhuthazayo, kungaba okujwayelekile.

• Ukufinyelela Emfundweni Yomhlaba Wonke: Abafundisi be-AI abazimele bangasiza ukufundisa abafundi ezindaweni ezinokushoda kothisha. Ithebhulethi enomfundisi we-AI ingase isebenze njengomfundisi oyinhloko wabafundi abanemfundo elinganiselwe, ehlanganisa ukufunda okuyisisekelo nezibalo. Ngo-2035, lokhu kungase kube okunye ukusetshenziswa okunomthelela omkhulu - i-AI yokuvala izikhala lapho othisha abangabantu bengatholakali. Kodwa-ke, ukuqinisekisa ikhwalithi nokufaneleka kwamasiko emfundo ye-AI ezimeni ezahlukene kuzobaluleka.

Ingabe i-AI izongena esikhundleni sothisha? Akunakwenzeka ngokuphelele. Ukufundisa kungaphezu kokuletha okuqukethwe - kuwukweluleka, ugqozi, ukwesekwa komphakathi-ngokomzwelo. Lezo zakhi zomuntu kunzima ukuthi i-AI ikwazi ukuziphindaphinda. Kodwa i-AI ingaba uthisha wesibili ekilasini noma ngisho uthisha wokuqala wokudlulisa ulwazi, ishiye othisha abangabantu bagxile kulokho abantu abakwenza kangcono kakhulu: ukuzwelana, ukugqugquzela, nokukhuthaza ukucabanga okujulile.

Kunokukhathazeka okufanele kuphathwe: ukuqinisekisa ukuthi i-AI inikeza ulwazi olunembile (akukho mibono efundisayo yamaqiniso angamanga), ukugwema ukuchema kokuqukethwe kwezemfundo, ukugcina ubumfihlo bedatha yabafundi, nokugcina abafundi bematasa (i-AI idinga ukugqugquzela, hhayi nje ukuthi ilungile). Cishe sizobona ukugunyazwa noma ukuqinisekiswa kwezinhlelo zemfundo ze-AI - okufana nezincwadi ezigunyazwayo - ukuqinisekisa ukuthi ziyahlangabezana nezindinganiso.

Enye inselele ukuthembela ngokweqile: uma umfundisi we-AI enikeza izimpendulo ngokushesha kakhulu, abafundi bangase bangafundi ukubekezela noma ukuxazulula izinkinga. Ukuze kuncishiswe lokhu, abafundisi be-AI besikhathi esizayo bangase bakhelwe ukuthi ngezinye izikhathi bavumele abafundi ukuba bazabalaze (njengoba umfundisi ongumuntu angaba) noma babakhuthaze ukuthi baxazulule izinkinga ngamasu kunokuba banikeze izixazululo.

Ngo-2035, ikilasi lingase liguqulwe: umfundi ngamunye onethuluzi elixhunywe nge-AI elibaqondisa ngejubane labo, kuyilapho uthisha ehlela imisebenzi yeqembu futhi enikeza ukuqonda komuntu. Imfundo ingase isebenze kahle futhi yenziwe ngendlela efanele. Isithembiso ukuthi wonke umfundi uthola usizo aludingayo lapho eludinga - ulwazi "lomfundisi womuntu siqu" lwangempela ngezikali. Ingozi iwukulahlekelwa ukuthintwa okuthile komuntu noma ukusebenzisa kabi i-AI (njengabafundi abakopela nge-AI). Kodwa-ke, uma ilawulwa kahle, i-AI ekhiqizayo imele ukwenza intando yeningi futhi ithuthukise ukufunda ngokuba umngane ohlale etholakala, onolwazi ohambweni lomfundi lwezemfundo.

I-Generative AI ku-Logistics and Supply Chain

Logistics - ubuciko nesayensi yezimpahla ezihambayo nokuphatha amaketango okunikezela - kungase kungabonakali njengesizinda esivamile se-AI "ekhiqizayo", kodwa ukuxazulula izinkinga nokuhlela okuhlakaniphile kubalulekile kulo mkhakha. I-Generative AI ingasiza ngokulingisa izimo, ukuthuthukisa izinhlelo, ngisho nokulawula amasistimu wamarobhothi. Umgomo wezokuthutha ukusebenza kahle nokonga izindleko, okuqondana kahle namandla e-AI ekuhlaziyeni idatha nokuphakamisa izixazululo. Ngakho-ke i-AI ingakwazi kanjani ukuzimela ekusebenziseni amaketanga wokuhlinzeka kanye nokusebenza kwempahla?

Amakhono Amanje (2025): Ukuthuthukisa kanye Nokuqondisa Ngokubona Komuntu

Namuhla, i-AI (kuhlanganise nezinye izindlela ezikhiqizayo) isetshenziswa kwezokuthutha ngokuyinhloko njengethuluzi lokusekela izinqumo :

• Ukuthuthukisa Umzila: Izinkampani ezifana ne-UPS ne-FedEx sezivele zisebenzisa ama-algorithms e-AI ukuze zandise imizila yokulethwa - iqinisekisa ukuthi abashayeli bathatha indlela esebenza kahle kakhulu. Ngokwesiko lawa bekungama-algorithms ocwaningo lwemisebenzi, kodwa manje izindlela ezikhiqizayo zingasiza ekuhloleni amanye amasu omzila ngaphansi kwezimo ezihlukahlukene (ithrafikhi, isimo sezulu). Ngenkathi i-AI iphakamisa imizila, abathumeli babantu noma abaphathi babeka imingcele (isb, izinto ezibalulekile) futhi bangabhala ngaphezulu uma kudingeka.

• Ukuhlelwa Kwemithwalo Nesikhala: Emalolini okupakisha noma iziqukathi zokuthutha, i-AI ingakhiqiza amasu okulayisha alungile (iliphi ibhokisi eliya lapho). I-AI ekhiqizayo ingase ikhiqize ukumiswa kokupakisha okuningi ukuze kwandiswe ukusetshenziswa kwesikhala, empeleni “ukudala” izixazululo abantu abangakhetha kuzo. Lokhu kwaqokonyiswa ucwaningo oluphawula ukuthi amaloli avame ukusebenzisa i-30% engenalutho e-US, futhi ukuhlela okungcono - okusizwa yi-AI - kunganciphisa leyo mfucuza ( Top Generative AI Use Cases in Logistics ). Lezi zinhlelo zokulayisha ezikhiqizwe yi-AI zihlose ukunciphisa izindleko zikaphethiloli kanye negesi ekhishwayo, futhi kwezinye izindawo zokugcina asetshenziswa ngokushintsha okuncane okwenziwa ngesandla.

• I-Demand Forecasting and Inventory Management: Amamodeli e-AI angabikezela ukufunwa komkhiqizo futhi akhiqize amasu okubuyisela isitoko. Imodeli ekhiqizayo ingase ilingise izimo zokufunwa okuhlukile (ake sithi, i-AI "icabanga" ukwanda kwesidingo ngenxa yeholide elizayo) futhi ihlele uhlu lwamagama ngokufanele. Lokhu kusiza abaphathi be-supply chain bazilungiselele. Njengamanje, i-AI inikeza izibikezelo neziphakamiso, kodwa abantu ngokuvamile benza ikholi yokugcina kumazinga okukhiqiza noma uku-oda.

• Ukuhlola Ubungozi: Uchungechunge lokuhlinzekwa kwempahla emhlabeni wonke lubhekene nokuphazamiseka (izinhlekelele zemvelo, ukubambezeleka kwechweba, izindaba zezepolitiki). Amasistimu e-AI manje ahlanganisa izindaba nedatha ukukhomba ubungozi emkhathizwe. Isibonelo, inkampani ethile yezokuthutha isebenzisa i-gen AI ukuze iskene i-inthanethi futhi ihlabe umkhosi amaphasishi ezinto zokuthutha ayingozi (izindawo okungenzeka zibe nenkinga ngenxa, ake sithi, isiphepho noma iziyaluyalu ezingenayo) ( I-Top Generative AI Use Cases in Logistics ). Ngalolo lwazi, abahleli bangakwazi ukuhlela kabusha ngokuzenzakalelayo ukuthunyelwa ezindaweni ezinenkinga. Kwezinye izimo, i-AI ingase income ngokuzenzakalelayo izinguquko zomzila noma izinguquko zezokuthutha, abantu bese bezivumela.

• I-Warehouse Automation: Izitolo eziningi zisebenza ngokuzenzakalelayo ngamarobhothi ukuze zicoshwe futhi zipakishwe. I-Generative AI ingakwazi ukwaba imisebenzi kumarobhothi nakubantu ukuze kugeleze kahle. Isibonelo, i-AI ingase ikhiqize ulayini wemisebenzi wabakhethi berobhothi njalo ekuseni ngokusekelwe kuma-oda. Lokhu kuvame ukuzimela ngokugcwele ekusebenzeni, nabaphathi baqapha nje ama-KPI - uma ama-oda ekhuphuka ngokungalindelekile, i-AI ilungisa ukusebenza iyodwa.

• Ukuphathwa Kwezimoto: I-AI isiza ukuhlela ukunakekelwa kwezimoto ngokuhlaziya amaphethini futhi ikhiqize amashejuli okulungisa afanele anciphisa isikhathi sokuphumula. Ingakwazi futhi ukuthunyelwa kweqembu ukuze kuncishiswe uhambo. Lezi zinqumo zingenziwa isoftware ye-AI ngokuzenzakalelayo inqobo nje uma ihlangabezana nezidingo zesevisi.

Sekukonke, kusukela ngo-2025, abantu babeka izinjongo (isb, “ukunciphisa izindleko kodwa uqinisekise ukulethwa kwezinsuku ezi-2”) futhi i-AI ikhipha izixazululo noma amashejuli ukuze kuzuzwe lokho. Amasistimu angasebenza usuku nosuku ngaphandle kokungenelela kuze kube yilapho kwenzeka okuthile okungajwayelekile. Izinto eziningi zokwenziwa kwezinto zibandakanya izinqumo eziphindaphindwayo (kufanele kuhambe nini lokhu kuthunyelwa? iyiphi inqolobane yokugcwalisa lo myalelo?), I-AI engafunda ukuyenza ngokungaguquki. Izinkampani zithembela kancane kancane i-AI ukuthi isingathe lezi zinqumo ezincane kanye nabaphathi abaqaphile kuphela lapho okuhlukile kwenzeka.

I-Outlook ka-2030-2035: Amaketanga Okuhlinzeka Ngokuzishayela

Eshumini leminyaka elizayo, singabona ngeso lengqondo ukusebenzelana okuzimele okuqhutshwa yi-AI:

• Izimoto Ezizihambelayo kanye nama-Drones: Amaloli azishayelayo nama-drones okulethwa, kuyilapho isihloko esibanzi se-AI/robhothi, sithinta ngokuqondile ukuhlelwa kwezinto. Ngo-2030, uma izinselelo zokulawula nezobuchwepheshe zinqotshwa, singase sibe namaloli okushayela i-AI emigwaqweni emikhulu noma ama-drones aphatha ukulethwa kwemayela lokugcina emadolobheni. Lawa ma-AI azokwenza izinqumo zesikhathi sangempela (izinguquko zemizila, ukugwema izithiyo) ngaphandle kwabashayeli abangabantu. I-engeli yokukhiqiza iwukuthi la ma-AI ezimoto afunda kanjani kudatha eningi nasezifanisweni, "ukuqeqeshwa" ngempumelelo ezimweni eziningi. Indiza ezimele ngokugcwele ingasebenza 24/7, abantu beqapha ukude kuphela. Lokhu kususa ingxenye enkulu yomuntu (abashayeli) emisebenzini yezokuthutha, okwandisa ukuzimela ngokumangazayo.

• Amaketanga Okuziphilisa Ngokuziphilisa: I-Generative AI cishe izosetshenziswa ukulingisa izimo ze-supply chain njalo futhi kulungiswe izinhlelo zezimo eziphuthumayo. Ngo-2035, i-AI ingase ibone ngokuzenzakalelayo lapho ifekthri yabahlinzeki bempahla ivaliwe (ngezindaba noma okuphakelayo kwedatha) futhi ngokushesha ishintshe ukutholakala kwabanye abahlinzeki esivele ibahlolile ekufaniseni. Lokhu kusho ukuthi i-supply chain "iyaziphilisa" ngokwayo ekuphazamisekeni lapho i-AI ithatha isinyathelo. Abaphathi abangabantu bazokwaziswa ngalokho okwenziwe yi-AI, kunalabo abaqala umsebenzi.

• I-End-to-End Inventory Optimization: I-AI ingaphatha ngokuzenzakalelayo i-inventory kuyo yonke inethiwekhi yezindawo zokugcina izimpahla nezitolo. Izonquma ukuthi isitokwe sihanjiswa nini futhi kuphi (mhlawumbe kusetshenziswa amarobhothi noma izimoto ezizihambelayo ukwenza kanjalo), sigcine inani elanele endaweni ngayinye. I-AI ngokuyisisekelo isebenzisa umbhoshongo wokulawula ukuthengwa kwempahla: ukubona konke ukugeleza nokwenza izinguquko ngesikhathi sangempela. Ngo-2035, umqondo wochungechunge lokuhlinzeka “ngokuzishayela ngokwakho” ungase usho ukuthi uhlelo lubala uhlelo olungcono kakhulu lokusabalalisa usuku ngalunye, ama-oda wemikhiqizo, amashejuli aqhutshwa yifekthri, futhi ahlele ezokuthutha zodwa. Abantu bebengamela isu jikelele futhi baphathe okuhlukile okungaphezu kokuqonda kwe-AI kwamanje.

• I-Generative Design ku-Logistics: Singabona i-AI iklama amanethiwekhi amasha ochungechunge lokunikezela. Ake sithi inkampani inwebela endaweni entsha; i-AI ingakhiqiza izindawo ezifanele zokugcina impahla, izixhumanisi zokuhamba, nezinqubomgomo zokusungula zaleso sifunda ngokunikezwa idatha - into eyenziwa ngabaxhumanisi nabahlaziyi namuhla. Ngo-2030, izinkampani zingase zithembele ezincomweni ze-AI zokukhethwa kwe-supply chain design, zithembe ukuthi izokala izinto ngokushesha futhi mhlawumbe zithole izixazululo zokudala (njengamahabhu okusabalalisa angabonakali) abantu abagejayo.

• Ukuhlanganiswa Nezokukhiqiza (Industry 4.0): Izinto zokuthutha azimi zodwa; ibophezela ekukhiqizeni. Amafekthri esikhathi esizayo angase abe nezinhlelo zokukhiqiza ze-AI ezikhiqizayo, zi-oda izinto ezingavuthiwe ngesikhathi, bese ziyala inethiwekhi yezokuthutha ukuthi ithumele imikhiqizo ngokushesha. Le AI edidiyelwe ingasho ukuhlelwa kwabantu okuncane kukonke - iketango elingenamthungo ukusuka ekwenziweni kuye ekulethweni eliqhutshwa ama-algorithms alungiselela izindleko, isivinini, nokusimama. Kakade, ngo-2025, amaketanga okuhlinzeka asebenza kahle aqhutshwa idatha; ngo-2035 kungenzeka kakhulu ukuthi ziqhutshwa yi-AI.

• Isevisi Yekhasimende Enamandla Kwezokuthutha: Ukwakhela phezu kwenkonzo yamakhasimende i-AI, ama-AI ochungechunge lokunikezela angase axhumane ngqo namakhasimende noma amaklayenti. Isibonelo, uma iklayenti elikhulu lifuna ukushintsha i-oda lalo lenqwaba ngomzuzu wokugcina, umenzeli we-AI angaxoxisana ngezinye izindlela ezingenzeka (njengokuthi “Singakwazi ukuletha uhhafu manje, uhhafu weviki elizayo ngenxa yezinkinga”) ngaphandle kokulinda umphathi ongumuntu. Lokhu kuhlanganisa ukuqonda kwe-AI yokukhiqiza zombili izinhlangothi (isidingo sekhasimende ngokumelene namandla okusebenza) kanye nokwenza izinqumo ezigcina ukusebenza kahle kuyilapho amakhasimende aneliseka.

Inzuzo elindelekile isistimu yezokuthutha esebenza kahle kakhudlwana, eqinile, futhi esabelayo Izinkampani zibona kusengaphambili ukonga okukhulu - uMcKinsey ulinganisela ukuthi ukulungiselelwa kochungechunge lokuhlinzekwa kwempahla oluqhutshwa yi-AI kunganciphisa kakhulu izindleko futhi kuthuthukise amazinga esevisi, kwengeze inani elingaba izigidigidi zenani ezimbonini zonkana ( Isimo se-AI ngo-2023: unyaka wokuphuma kwe-Generative AI | McKinsey ).

Kodwa-ke, ukuguqula ukulawula okwengeziwe ku-AI nakho kunezingozi, njengamaphutha e-cascading uma ingqondo ye-AI inephutha (isb, isimo esidumile sochungechunge lokunikezela nge-AI olukhipha inkampani ngokungazi ngenxa yephutha lokumodela). Izivikelo ezifana nokuthi "umuntu-ke-loop ezinqumweni ezinkulu" noma okungenani amadeshibhodi avumela ukukhishwa ngokushesha kwabantu cishe azohlala kuze kube u-2035. Ngokuhamba kwesikhathi, njengoba izinqumo ze-AI zibonisa, abantu bazokhululeka kakhudlwana ukubuyela emuva.

Kuyathakazelisa ukuthi ngokulungiselela ukusebenza kahle, i-AI kwesinye isikhathi ingenza izinqumo ezingqubuzana nokuthandwa ngabantu noma imikhuba yendabuko. Isibonelo, ukwenza kahle kakhulu kungase kuholele kuhlu lwempahla oluncane kakhulu, olusebenza kahle kodwa olungazizwa luyingozi. Ochwepheshe be-Supply chain ngo-2030 kungase kudingeke balungise imibono yabo ngoba i-AI, eqoqa idatha enkulu, ingase ibonise ukuthi isu layo elingajwayelekile empeleni lisebenza kangcono.

Okokugcina, kufanele sicabangele ukuthi izithiyo ezibonakalayo (ingqalasizinda, isivinini senqubo ebonakalayo) zikhawulela ukuthi ukuhlelwa kwezinto kungashintsha ngokushesha kangakanani, ngakho uguquko lapha lumayelana nokuhlela okuhlakaniphile nokusetshenziswa kwezimpahla kunokuthi kube yinto engokoqobo entsha ngokuphelele. Kepha nangaphakathi kwaleyo mingcele, izixazululo zokudala ze-AI kanye nokwenza kahle okungaphezi kungase kuthuthukise ngendlela emangalisayo ukuthi izimpahla zihamba kanjani emhlabeni ngokuhlela okuncane okwenziwa ngesandla.

Kafushane, izinto zokusebenza ngo-2035 zingase zisebenze ngokufana nomshini ozishintshayo ogcotshwe kahle: izimpahla ezigeleza kahle, imizila elungiswa ngesikhathi sangempela ukuze kube nokuphazamiseka, izindawo zokugcina izinto eziziphatha ngamarobhothi, kanye nalo lonke uhlelo luqhubeka nokufunda futhi luthuthuke kudatha - konke kuhlelwa yi-generative AI esebenza njengobuchopho bomsebenzi.

I-Generative AI Kwezezimali Nezamabhizinisi

Imboni yezezimali isebenzisana kakhulu nolwazi - imibiko, ukuhlaziya, ukuxhumana namakhasimende - okwenza kube yindawo evundile ye-AI ekhiqizayo. Kusukela ebhange kuya ekuphathweni kokutshalwa kwezimali kanye nomshwalense, izinhlangano zihlola i-AI ukuze zenze i-automation kanye nokukhiqiza ukuqonda. Umbuzo uwukuthi, yimiphi imisebenzi yezezimali i-AI engasingatha ngokwethembeka ngaphandle kokuqondisa komuntu, uma kubhekwa ukubaluleka kokunemba nokwethenjwa kulesi sizinda?

Amandla Amanje (2025): Imibiko Ezenzakalelayo Nosekelo Lwezinqumo

Kusukela namuhla, i-AI ekhiqizayo inikela kwezezimali ngezindlela ezimbalwa, imvamisa ngaphansi kokuqondisa komuntu:

• Isizukulwane Sombiko: Amabhange namafemu ezezimali akhiqiza imibiko eminingi - izifinyezo zeholo, ukuphawula kwemakethe, ukuhlaziya iphothifoliyo, njll. I-AI isivele isetshenziselwa ukubhala lokhu. Isibonelo, i-Bloomberg ithuthukise i-BloombergGPT , imodeli yolimi enkulu eqeqeshelwe idatha yezezimali, ukuze isize ngemisebenzi efana nokuhlukaniswa kwezindaba kanye ne-Q&A yabasebenzisi bayo betheminali ( I-Generative AI iza ngezimali ). Nakuba ukusetshenziswa kwayo okuyinhloko kuwukusiza abantu ukuthi bathole ulwazi, kukhombisa indima ekhulayo ye-AI. I-Automated Insights (inkampani i-AP esebenze nayo) nayo ikhiqize ama-athikili ezezimali. Izincwadi zezindaba zokutshalwa kwezimali eziningi zisebenzisa i-AI ukuphinda umnyakazo wemakethe wansuku zonke noma izinkomba zezomnotho. Ngokuvamile, abantu babuyekeza lokhu ngaphambi kokukuthumela kumakhasimende, kodwa ukuhlela okusheshayo kunokubhala kusukela ekuqaleni.

• Ukuxhumana Kwekhasimende: Emabhange okudayisa, izingxoxo ze-AI zisingatha imibuzo yamakhasimende mayelana namabhalansi e-akhawunti, ukuthengiselana, noma ulwazi lomkhiqizo (ukuhlanganisa esizindeni sesevisi yamakhasimende). Futhi, i-AI ingakhiqiza izincwadi zezeluleko zezezimali ezenzelwe umuntu siqu noma ama-nudges. Isibonelo, i-AI ingase ihlonze ukuthi ikhasimende lingonga ezinkokhelweni futhi libhale ngokuzenzakalelayo umlayezo ophakamisa ukuthi lishintshele ohlotsheni oluhlukile lwe-akhawunti, bese luphuma ngaphandle kokungenelela kwabantu okuncane. Lolu hlobo lokuxhumana komuntu siqu esikalini luwukusetshenziswa kwamanje kwe-AI kwezezimali.

• Ukutholwa Nezaziso Zokukhwabanisa: I-Generative AI ingasiza ekudaleni ukulandisa noma izincazelo zokudidayo okutholwe amasistimu okukhwabanisa. Isibonelo, uma umsebenzi osolisayo umakiwe, i-AI ingase ikhiqizele ikhasimende umlayezo wencazelo (“Siqaphele ukungena ngemvume kusuka kudivayisi entsha…”) noma umbiko wabahlaziyi. Ukutholwa kuyazenzakalela (kusetshenziswa ukutholwa okudidayo kwe-AI/ML), futhi ukuxhumana kuya ngokuya kuzenzakalela, nakuba izenzo zokugcina (ukuvimbela i-akhawunti) kuvame ukuba nokuhlola okuthile komuntu.

• Ukwelulekwa Ngezezimali (okunomkhawulo): Abanye abeluleki be-robo (izinkundla zokutshala imali ezizenzakalelayo) basebenzisa ama-algorithms (hhayi ngempela i-AI ekhiqizayo) ukuze baphathe amaphothifoliyo abangenabo abeluleki abangabantu. I-Generative AI ingena ngokuthi, ake sithi, ikhiqiza ukuphawula kokuthi kungani ukuhweba okuthile kwenziwa noma isifinyezo sokusebenza kwephothifoliyo eklanyelwe iklayenti. Nokho, iseluleko esimsulwa sezezimali (njengokuhlelwa kwezimali okuyinkimbinkimbi) kuseyi-algorithmic evamile yomuntu noma esekelwe emithethweni; iseluleko sokukhiqiza samahhala ngaphandle kokugadwa siyingozi ngenxa yesibopho sezomthetho uma singalungile.

• Ukuhlolwa Kwengozi Nokubhala Ngaphansi: Izinkampani zomshwalense zihlola i-AI ukuze zibhale ngokuzenzakalelayo imibiko yokuhlola ubungozi noma imibhalo yenqubomgomo esalungiswa. Isibonelo, uma kunikezwe idatha emayelana nesakhiwo, i-AI ingakhiqiza inqubomgomo yomshwalense esalungiswa noma umbiko wombhali ochaza izinto eziyingozi. Abantu okwamanje babuyekeza le miphumela ngoba noma yiliphi iphutha kunkontileka lingaba yizindleko ezinkulu.

• Ukuhlaziywa Kwedatha Nemininingwane: I-AI ingasebenzisa izitatimende zezimali noma izindaba futhi ikhiqize izifinyezo. Abahlaziyi basebenzisa amathuluzi angakwazi ukufingqa ngokushesha umbiko wonyaka onamakhasi ayi-100 abe amaphuzu abalulekile, noma akhiphe okuzothathwa okuyinhloko kokulotshiweyo kwekholi yezinzuzo. Lezi zifinyezo zonga isikhathi futhi zingasetshenziswa ngokuqondile ekuthathweni kwezinqumo noma ezidluliswayo, kodwa abahlaziyi abahlakaniphile bahlola kabili imininingwane ebalulekile.

Empeleni, i-AI yamanje kwezezimali isebenza njengomhlaziyi/umbhali ongakhathali , ekhiqiza okuqukethwe okupholishwa abantu. Ukusetshenziswa okuzimele ngokugcwele kusezindaweni ezichazwe kahle njengezindaba ezishayelwa yidatha (akukho ukwahlulela okuqondile okudingekayo) noma izimpendulo zesevisi yamakhasimende. Ukwethemba ngokuqondile i-AI ngezinqumo mayelana nemali (njengezimali ezihambayo, ukwenza ukuhweba okungaphezu kwama-algorithms asethiwe) akuvamile ngenxa yezibalo eziphezulu kanye nokuhlolwa kokulawula.

I-Outlook ka-2030-2035: Abahlaziyi be-AI kanye Nokusebenza Kwezezimali Okuzenzakalelayo

Uma sibheka phambili, ngo-2035 i-AI ekhiqizayo ingase igxiliswe ngokujulile ekusebenzeni kwezezimali, okungenzeka isingathe imisebenzi eminingi ngokuzimela:

• Abahlaziyi Bezezimali be-AI: Singabona izinhlelo ze-AI ezingahlaziya izinkampani nezimakethe futhi zikhiqize izincomo noma imibiko ezingeni lomhlaziyi wocwaningo ngokulingana kwabantu. Ngo-2030, i-AI ingase ifunde konke okugciniwe kwezezimali yenkampani, iqhathanise nedatha yomkhakha, futhi ikhiqize umbiko wezincomo zokutshalwa kwezimali (“Thenga/Thengisa” ngokucabanga) iyodwa. Ezinye izimali ze-hedge sezivele zisebenzisa i-AI ukukhiqiza izimpawu zokuhweba; ngawo-2030, imibiko yocwaningo lwe-AI ingase ibe yinto evamile. Abaphathi bephothifoliyo yabantu bangase baqale ukwethemba ukuhlaziya okukhiqizwa yi-AI njengokufaka okukodwa phakathi kokunye. Kukhona namathuba okuthi i-AI ilawule ngokuzimele iphothifoliyo: ngokuqhubekayo iqaphe futhi ilinganise kabusha ukutshalwa kwezimali ngokwesu elichazwe ngaphambilini. Eqinisweni, ukuhweba nge-algorithmic sekuvele kuzenzakalelayo kakhulu - i-AI ekhiqizayo ingase yenze amasu avumelane nezimo ngokukhiqiza nokuhlola amamodeli amasha okuhweba ngokwawo.

• Ukuhlela Kwezezimali Okuzenzakalelayo: Abeluleki be-AI ababhekene nabathengi bangakwazi ukusingatha ukuhlelwa kwezimali okujwayelekile kwabantu. Ngo-2030, ungase utshele i-AI izinhloso zakho (ukuthenga indlu, ukongela ikolishi) futhi ingakhiqiza uhlelo oluphelele lwezezimali (isabelomali, ukwabiwa kokutshalwa kwezimali, iziphakamiso zomshwalense) ezenzelwe wena. Ekuqaleni umhleli wezezimali ongumuntu angase ayibuyekeze, kodwa njengoba ukuzethemba kukhula, iseluleko esinjalo singanikezwa ngokuqondile kubathengi, ngokuzihlangula okufanele. Ukhiye kuzoba ukuqinisekisa ukuthi izeluleko ze-AI ziyahambisana nemithetho futhi zizuzisa ikhasimende. Uma ixazululwa, i-AI ingenza izeluleko eziyisisekelo zezezimali zifinyeleleke kalula ngezindleko eziphansi.

• I-Back-Office Automation: I-Generative AI ingase iphathe ngokuzimele imibhalo eminingi yasehhovisi - izicelo zemalimboleko, imibiko yokuthobela imithetho, izifinyezo zocwaningo. Isibonelo, i-AI ingathatha yonke idatha yokwenziwayo futhi ikhiqize umbiko wokucwaninga ohlaba umkhosi noma yikuphi ukukhathazeka. Abacwaningi mabhuku ngo-2035 bangase bachithe isikhathi esiningi bebuyekeza okuhlukile okuhlatshwe umkhosi kwe-AI kunokuba bahlanganise yonke into ngokwabo. Ngokufanayo, ukuze kuthotshelwe umthetho, i-AI ingakhiqizela abalawuli imibiko yemisebenzi esolisayo (ama-SAR) ngaphandle kokuba umhlaziyi ayibhale kusukela ekuqaleni. Ukukhiqiza ngokuzenzakalelayo lawa madokhumenti ajwayelekile, lapho ukugadwa komuntu kuya ngendlela ehlukile, kungase kube okujwayelekile.

• Izicelo Zomshwalense Nokubhala Ngaphansi: I-AI ingacubungula isimangalo somshwalense (ngobufakazi bezithombe, njll.), inqume ukukhava, futhi ikhiqize ngokuzenzakalelayo incwadi yesinqumo senkokhelo. Singase sifinyelele iphuzu lapho izimangalo eziqondile (njengezingozi zezimoto ezinedatha ecacile) zixazululwa ngokuphelele yi-AI phakathi nemizuzu yokuthunyelwa. Ukubhala ngaphansi kwezinqubomgomo ezintsha kungase kufane: I-AI ihlola ubungozi futhi ikhiqize imigomo yenqubomgomo. Ngo-2035, mhlawumbe kuphela amacala ayinkimbinkimbi noma asemnceleni afinyelela kubantu ababhala phansi.

• Ukukhwabanisa Nokuphepha: I-AI cishe izoba bucayi kakhulu ekutholeni nasekuphenduleni ukukhwabanisa noma izinsongo ze-inthanethi kwezezimali. Abenzeli be-AI abazimele bangase baqaphe okwenziwayo ngesikhathi sangempela futhi bathathe izinyathelo ezisheshayo (ukuvimba ama-akhawunti, ukumisa ukuthengiselana) lapho imibandela ethile ifika, bese beveza isizathu. Isivinini sibalulekile lapha, ngakho-ke ukubandakanyeka okuncane komuntu kuyadingeka. Ingxenye ekhiqizayo ingase ingene ekukhulumeni lezi zenzo kumakhasimende noma abalawuli ngendlela ecacile.

• Ukwesekwa Kwabaphathi: Cabanga “ngesikhulu sabasebenzi” se-AI esingakhiqizela abaphathi imibiko yebhizinisi ngokushesha. Buza, "Ingabe uphiko lwethu lwase-Europe lusebenze kanjani kule kota futhi babeyini abashayeli abakhulu uma kuqhathaniswa nonyaka odlule?" futhi i-AI izokhiqiza umbiko omfushane onamashadi, wonke anembile, adonsa kudatha. Lolu hlobo lokubika olushintshashintshayo, oluzimele nokuhlaziya lungaba lula njengengxoxo. Ngo-2030, ukubuza i-AI ngobuhlakani bebhizinisi nokuyethemba ukuthi izonikeza izimpendulo ezifanele kungase kuthathe indawo yemibiko emile futhi mhlawumbe nezindima zomhlaziyi.

Ukuqagela okukodwa okuthakazelisayo: ngeminyaka yawo-2030, iningi lokuqukethwe kwezezimali (izindaba, imibiko, njll.) kungenzeka kube kukhiqizwe i-AI . Kakade, izitolo ezifana ne-Dow Jones ne-Reuters zisebenzisa okuzenzakalelayo kwezinye izingcezu zezindaba. Uma lowo mkhuba uqhubeka, futhi uma kubhekwa ukuqhuma kwedatha yezezimali, i-AI ingase ibe nesibopho sokuhlunga nokuxhumana iningi layo.

Nokho, ukwethenjwa nokuqinisekisa kuzoba okuphakathi. Imboni yezezimali ilawulwa kakhulu futhi noma iyiphi i-AI esebenza ngokuzenzakalelayo izodinga ukuhlangabezana namazinga aqinile:

• Ukuqinisekisa ukuthi awekho ama-allucinations (awukwazi ukuba nomhlaziyi we-AI osungula imethrikhi yezezimali okungelona iqiniso - engadukisa izimakethe).

• Ukugwema ukwenzelela noma imikhuba engekho emthethweni (njengokuphinda wenze izinqumo zokuboleka ungahlosile ngenxa yedatha yokuqeqeshwa echemile).

• Ukucwaningwa kwamabhuku: abalawuli bazodinga ukuthi izinqumo ze-AI zichazwe. Uma i-AI yenqaba ukubolekwa imali noma yenza isinqumo sokuhweba, kufanele kube nesizathu esingahlolwa. Amamodeli akhiqizayo angaba yibhokisi elimnyama, ngakho-ke lindela ukuthuthukiswa kwamasu e-AI achazekayo ukwenza izinqumo zawo zibe sobala.

Iminyaka eyi-10 ezayo cishe izobandakanya ukubambisana okuseduze phakathi kwe-AI nochwepheshe bezezimali, kancane kancane inyakaze umugqa wokuzimela njengoba ukuzethemba kukhula. Ukuwina kwangaphambi kwesikhathi kuzoza nge-automation enengcuphe ephansi (njengokwenza umbiko). Okunzima kuzoba yizinqumo ezibalulekile njengezinqumo zezikweletu noma ukukhetha ukutshalwa kwezimali, kodwa nalapho, njengoba irekhodi le-AI likhula, amafemu angase ayinikeze ukuzimela okwengeziwe. Isibonelo, mhlawumbe isikhwama se-AI sizosebenza nombonisi ongumuntu ongenelela kuphela uma ukusebenza kwehluka noma uma i-AI ihlaba umkhosi ukungaqiniseki.

Ngokomnotho, uMcKinsey ulinganisele ukuthi i-AI (ikakhulukazi i-gen AI) ingangeza nge-oda lama-dollar ayizigidi eziyizinkulungwane ezingama-200-340 enanini lokubhanga minyaka yonke kanye nemithelela emikhulu efanayo kumshwalense nezimakethe ezinkulu ( Isimo se-AI ngo-2023: unyaka wokuphuma kwe-Generative AI | McKinsey ) ( Liyini ikusasa le-Generative AI ? Lokhu kungenxa yempumelelo kanye nemiphumela yezinqumo ezingcono. Ukuze uthwebule lelo nani, ukuhlaziya okuningi okujwayelekile kwezezimali kanye nokuxhumana cishe kuzodluliselwa kumasistimu e-AI.

Kafushane, ngo-2035 i-AI ekhiqizayo ingase ifane nebutho labahlaziyi abancane, abeluleki, nomabhalane abasebenza kuwo wonke umkhakha wezezimali, benza umsebenzi omningi we-grunt kanye nokuhlaziya okuthile okuyinkimbinkimbi ngokuzimele. Abantu basazobeka imigomo futhi baphathe isu lezinga eliphezulu, ubudlelwano bamakhasimende, nokugada. Umhlaba wezezimali, ngokuqaphela, uzonweba ukuzimela kancane kancane - kodwa isiqondiso sisobala ukuthi ukwanda kokucubungula ulwazi kanye nezincomo zesinqumo kuzovela ku-AI. Ngokufanelekile, lokhu kuholela kusevisi esheshayo (imalimboleko esheshayo, iseluleko samahora nehora), izindleko eziphansi, kanye nokungaba namandla okukhulu (izinqumo ezisekelwe kumaphethini wedatha). Kodwa ukulondoloza ukwethembana kuyoba okubalulekile; iphutha elilodwa lephrofayili ephezulu ye-AI kwezezimali lingabangela umonakalo omkhulu (cabanga ukuphahlazeka kwe-flash okucushwe yi-AI noma inzuzo enqatshelwe ngokungalungile ezinkulungwaneni zabantu). Ngakho-ke, ama-guardrails nokuhlolwa kwabantu kungenzeka kuqhubeke ikakhulukazi ezenzweni ezibhekene nabathengi, njengoba nezinqubo zasehhovisi zisebenza ngokukhululeka.

Izinselelo Nokucatshangelwa Kokuziphatha

Kuzo zonke lezi zizinda, njengoba i-AI ekhiqizayo ithatha izibopho ezizimele, isethi yezinselelo ezivamile nemibuzo yokuziphatha iyaphakama. Ukuqinisekisa ukuthi i-AI iyi-ejenti ezimele ethembekile nezuzisayo akuwona nje umsebenzi wezobuchwepheshe, kodwa owomphakathi. Lapha siveza izinto ezikhathazayo ezibalulekile nokuthi zisingathwa kanjani (noma ezizodinga ukubhekwana nazo):

Ukwethembeka Nokunemba

Inkinga Ye-Hallucination: Amamodeli e-Generative AI angakhiqiza imiphumela engalungile noma eyenziwe ngokuphelele ebukeka izethemba. Lokhu kuyingozi ikakhulukazi lapho kungekho muntu obambe iqhaza ekubambeni amaphutha. I-chatbot ingase inikeze ikhasimende imiyalelo engalungile, noma umbiko obhalwe nge-AI ungase uqukathe izibalo ezenziwe. Kusukela ngo-2025, ukunganembi kubonwa njengengcuphe ephezulu ye-AI ekhiqizayo yizinhlangano ( Isimo se-AI ngo-2023: unyaka wokuphuma kwe-Generative AI | McKinsey ) ( Isimo se-AI: inhlolovo yomhlaba wonke | McKinsey ). Ukuqhubekela phambili, amasu afana nokuhlola amaqiniso ngokumelene nesizindalwazi, ukuthuthukiswa kwemodeli yezakhiwo, nokufunda okuqiniswayo ngempendulo kuthunyelwa ukuze kuncishiswe imibono engekho. Izinhlelo ezizenzakalelayo ze-AI cishe zizodinga ukuhlolwa okuqinile futhi mhlawumbe nokuqinisekiswa okusemthethweni kwemisebenzi ebalulekile (njengokwenza ikhodi okungase kwethule iziphazamisi/amaphutha okuvikela uma kungalungile).

Ukungaguquguquki: Amasistimu e-AI adinga ukusebenza ngokuthembekile ngokuhamba kwesikhathi nakuzo zonke izimo. Isibonelo, i-AI ingase yenze kahle emibuzweni evamile kodwa ikhubeke emacaleni asemaphethelweni. Ukuqinisekisa ukusebenza okungaguquguquki kuzodinga idatha yokuqeqeshwa ebanzi ehlanganisa izimo ezihlukahlukene nokuqapha okuqhubekayo. Izinhlangano eziningi zihlela ukuba nezindlela ezixubile - i-AI iyasebenza, kodwa amasampula angahleliwe acwaningwa ngabantu - ukukala amanani okunemba aqhubekayo.

I-Fail-Safes: Uma i-AI izimele, ukukwazi ukubona ukungaqiniseki kwayo kubalulekile. Isistimu kufanele yakhelwe “ukwazi uma ingazi.” Isibonelo, uma udokotela we-AI engenaso isiqiniseko sokuxilongwa, kufanele amake ukuze kubuyekezwe umuntu kunokuba anikeze ukuqagela okungahleliwe. Ukwakha isilinganiso sokungaqiniseki emiphumeleni ye-AI (kanye nokuba nemikhawulo yokunikezwa komuntu okuzenzakalelayo) kuyindawo esebenzayo yentuthuko.

Ukuchema Nobulungiswa

I-Generative AI ifunda kudatha yomlando engaqukatha ukuchema (ubuhlanga, ubulili, njll.). I-AI ezimele ingase iqhubekisele phambili noma ikhulise lokho kubandlulula:

• Ekuqashweni noma ekwamukelweni, umenzi wesinqumo we-AI angabandlulula ngokungafanele uma idatha yakhe yokuqeqeshwa inokuchema.

• Kusevisi yamakhasimende, i-AI ingase iphendule ngendlela ehlukile kubasebenzisi ngokusekelwe kulimi lwesigodi noma ezinye izici ngaphandle kokuthi ihlolwe ngokucophelela.

• Ezindaweni zokudala, i-AI ingase imele kancane amasiko noma izitayela ezithile uma isethi yokuqeqeshwa ibingalingani.

Ukubhekana nalokhu kudinga ukucutshungulwa kwedathasethi ngokucophelela, ukuhlolwa kokuchema, mhlawumbe nokulungiswa kwe-algorithmic ukuze kuqinisekiswe ubulungisa. Ukungafihli kusemqoka: izinkampani zizodinga ukudalula imibandela yesinqumo se-AI, ikakhulukazi uma i-AI ezimele ithinta amathuba noma amalungelo othile (njengokuthola imali mboleko noma umsebenzi). Abalawuli sebevele banakile; isb, Umthetho we-EU we-AI (emisebenzini kusukela maphakathi nawo-2020s) uzodinga ukuhlolwa kokuchema kumasistimu e-AI ayingozi kakhulu.

Ukuziphendulela kanye Nesibopho Sezomthetho

Uma uhlelo lwe-AI olusebenza ngokuzenzakalelayo lulimaza noma lenza iphutha, ubani onesibopho? Izinhlaka zomthetho ziyaqhubeka:

• Izinkampani ezithumela i-AI cishe zizobamba isikweletu, okufana nokuba nesibopho ngezenzo zesisebenzi. Isibonelo, uma i-AI inikeza iseluleko esibi sezezimali esiholela ekulahlekelweni, inkampani kungase kudingeke ukuthi inxephezele iklayenti.

• Kunenkulumompikiswano mayelana ne-AI “ubuntu” noma ukuthi i-AI ethuthukisiwe ingaba necala ngokwengxenye yini, kodwa lokho kunengqondo kakhulu manje. Empeleni, ukusolwa kuzobuyela emuva konjiniyela noma opharetha.

• Imikhiqizo emisha yomshuwalense ingase ivele ngenxa yokwehluleka kwe-AI. Uma iloli elizishayelayo lidala ingozi, umshwalense womkhiqizi ungase uyikhave, okufana nesibopho somkhiqizo.

• Ukubhalwa phansi kanye nokugawulwa kwezinqumo ze-AI kuzobaluleka ekuhlolweni kwezidumbu. Uma kukhona okungahambi kahle, sidinga ukuhlola umkhondo wesinqumo se-AI ukuze sifunde kuso futhi sabelane nesibopho. Abalawuli bangase bagunyaze ukugawulwa kwemithi ngezenzo ze-AI ezizimele ngenxa yalesi sizathu.

Ukungafihli nokuchazwa

I-Autonomous AI kufanele ikwazi ukuchaza ukucabanga kwayo ngamagama aqondakalayo abantu, ikakhulukazi emikhakheni elandelanayo (ezezimali, ukunakekelwa kwezempilo, uhlelo lwezobulungiswa). I-AI echazekayo inkambu elwela ukuvula ibhokisi elimnyama:

• Ngokunqatshelwa kwemalimboleko yi-AI, imithethonqubo (njengase-US, ECOA) ingase idinge ukunikeza umfakisicelo isizathu. Ngakho-ke i-AI kufanele ikhiphe izici (isb, "isilinganiso esiphezulu sesikweletu-kuya-imali engenayo") njengencazelo.

• Abasebenzisi abasebenzisana ne-AI (njengabafundi abanomfundisi we-AI noma iziguli ezinohlelo lokusebenza lwezempilo lwe-AI) bafanelwe ukwazi ukuthi zifika kanjani ezilulekweni. Imizamo iyaqhubeka yokwenza ukucabanga kwe-AI kulandeleke kakhudlwana, ngokwenza amamodeli alula noma ngokuba namamodeli achazayo afanayo.

• Ukubonisa ngale kuphinde kusho ukuthi abasebenzisi kufanele bazi uma bebhekene ne-AI uma beqhathaniswa nomuntu. Imihlahlandlela yokuziphatha (futhi mhlawumbe nemithetho ethile) incike ekufuneni ukudalulwa uma ikhasimende likhuluma ne-bot. Lokhu kuvimbela ukukhohlisa futhi kuvumela imvume yomsebenzisi. Ezinye izinkampani manje zimaka ngokucacile okuqukethwe okubhalwe i-AI (njengokuthi “Le ndatshana yenziwe i-AI”) ukuze kugcinwe ukwethenjwa.

Ubumfihlo Nokuvikelwa Kwedatha

I-Generative AI ivamise ukudinga idatha - okuhlanganisa idatha yomuntu siqu okungenzeka ibe bucayi - ukuze isebenze noma ifunde. Imisebenzi ezenzakalelayo kufanele ihloniphe ubumfihlo:

• I-ejenti yesevisi yekhasimende ye-AI izofinyelela ulwazi lwe-akhawunti ukuze isize ikhasimende; leyo datha kufanele ivikelwe futhi isetshenziselwe umsebenzi kuphela.

• Uma abafundisi be-AI bekwazi ukufinyelela kumaphrofayela abafundi, kunokucatshangwa ngaphansi kwemithetho efana ne-FERPA (e-US) ukuze kuqinisekiswe ubumfihlo bedatha yezemfundo.

• Amamodeli amakhulu angakhumbula ngokungaqondile imininingwane yawo yokuqeqeshwa (isb, ukuphindaphinda ikheli lomuntu elibonwa ngesikhathi sokuqeqeshwa). Amasu afana nokwahluka kobumfihlo kanye nokwenza idatha ingaziwa ekuqeqeshweni kubalulekile ukuze kuvinjelwe ukuvuza kolwazi lomuntu siqu emiphumeleni ekhiqiziwe.

• Imithetho efana ne-GDPR inika abantu amalungelo ngezinqumo ezizenzakalelayo ezibathintayo. Abantu bangacela ukubuyekezwa komuntu noma izinqumo ukuthi zingazenzeki kuphela uma ziba nomthelela omkhulu kuzo. Ngo-2030, le mithetho ingase iguquke njengoba i-AI idlanga kakhulu, ngokunokwenzeka yethule amalungelo encazelo noma ukuphuma ekucubungulweni kwe-AI.

Ukuphepha Nokuhlukumeza

Amasistimu e-AI azimele angaba okuhloswe ngawo ukugetshengwa noma angaxhashazwa ukwenza izinto ezinonya:

• Ijeneretha yokuqukethwe ye-AI ingasetshenziswa kabi ukuze kudaleke i-disinformation esikalini (amavidiyo omgunyathi, izindatshana zezindaba ezingamanga), okuyingozi emphakathini. Ukuziphatha kokukhipha amamodeli akhiqizayo anamandla kuphikiswana ngakho kakhulu (i-OpenAI ekuqaleni ibiqaphile ngamakhono esithombe se-GPT-4, ngokwesibonelo). Izixazululo zihlanganisa okuqukethwe okukhiqizwe yi-AI nge-watermarking ukusiza ukuthola amanga, nokusebenzisa i-AI ukulwa ne-AI (njengama-algorithms okuthola ama-deepfakes).

• Uma i-AI ilawula izinqubo zomzimba (ama-drones, izimoto, ukulawula kwezimboni), ukuyivikela ngokumelene nokuhlasela kwe-cyber kubalulekile. Isistimu yokuzimela entshontshiwe ingabangela ukulimala komhlaba wangempela. Lokhu kusho ukubethela okuqinile, ukuhluleka kokuphepha, kanye nekhono lokubhala ngaphezulu komuntu noma ukuvala shaqa uma okuthile kubonakala kusengozini.

• Kuphinde kube nokukhathazeka kokuthi i-AI idlula imingcele ehlosiwe (isimo “se-AI esinamandla”). Nakuba ama-AI amanje engenaso i-ejensi noma inhloso, uma amasistimu okuzimela esikhathi esizayo esebenza kakhudlwana, kudingeka imingcele eqinile nokuqapha ukuze kuqinisekiswe ukuthi awenzi, akenzi ukuhweba okungagunyaziwe noma aphule imithetho ngenxa yenhloso engashiwongo kahle.

Ukusetshenziswa Kwezimiso Zokuhle kanye Nomphumela Womuntu

Okokugcina, ukucatshangelwa okubanzi kokuziphatha:

• Ukususwa Kwemisebenzi: Uma i-AI ingenza imisebenzi ngaphandle kokungenelela komuntu, kwenzekani kuleyo misebenzi? Ngokomlando, ubuchwepheshe buzenzela eminye imisebenzi kodwa budale eminye. Inguquko ingaba buhlungu kubasebenzi amakhono abo asezindaweni ezizenzekelayo. Umphakathi uzodinga ukuphatha lokhu ngokusebenzisa amakhono kabusha, imfundo, kanye nokucabanga kabusha ngosekelo lwezomnotho (abanye baphakamisa ukuthi i-AI ingase idinge imibono efana neholo eliyisisekelo jikelele uma umsebenzi omningi uzenzakalela). Kakade, izinhlolovo zikhombisa imizwa exubile - ucwaningo olulodwa lwathola ingxenye yesithathu yabasebenzi ikhathazekile ngokushintsha kwe-AI esikhundleni, kanti abanye bakubona njengokuthatha isihluku.

• Ukuguguleka Kwamakhono Obuntu: Uma abafundisi be-AI befundisa futhi i-AI autopilots ishayela futhi i-AI ibhala ikhodi, ingabe abantu bazolahlekelwa yilawa makhono? Ukuthembela ngokweqile ku-AI esimweni esibi kakhulu kungase kucekele phansi ubuchwepheshe; yinto ezodinga ukuthi izinhlelo zemfundo nokuqeqesha zilungiswe, kuqinisekiswe ukuthi abantu basafunda okuyisisekelo ngisho noma i-AI isiza.

• Ukwenza Izinqumo Zokuziphatha: I-AI ayinakho ukwahlulela kokuziphatha komuntu. Kwezokunakekelwa kwempilo noma umthetho, izinqumo ezitholwe yidatha kuphela zingase zingqubuzane nozwelo noma ubulungisa ezimeni ezingazodwana. Kungase kudingeke ukuthi sihlanganise izinhlaka zezimiso zokuziphatha ku-AI (indawo yocwaningo lwezimiso ze-AI, isb, ukuqondisa izinqumo ze-AI nezindinganiso zomuntu). Okungenani, ukugcina abantu benolwazi mayelana nezinqumo ezihambisana nezimiso zokuziphatha kuhle.

• Ukubandakanywa: Ukuqinisekisa ukuthi izinzuzo ze-AI zisatshalaliswa kabanzi kuwumgomo wesimilo. Uma kuphela izinkampani ezinkulu zingakwazi ukukhokhela i-AI ethuthukisiwe, amabhizinisi amancane noma izifunda ezimpofu zingasala ngemuva. Imizamo yomthombo ovulekile nezisombululo ze-AI ezithengekayo zingasiza ukwenza intando yeningi ukufinyelela. Futhi, izindawo zokusebenzelana kufanele zakhelwe ukuze noma ubani akwazi ukusebenzisa amathuluzi e-AI (izilimi ezahlukene, ukufinyeleleka kwalabo abakhubazekile, njll.), funa sidale ukuhlukana okusha kwedijithali kokuthi “ubani onomsizi we-AI futhi ongenalo.”

Ukunciphisa Ubungozi Kwamanje: Ngasohlangothini oluhle, njengoba izinkampani zikhipha i-gen AI, kukhona ukuqwashisa okukhulayo kanye nezindlela kulezi zinkinga. Ngasekupheleni kuka-2023, cishe uhhafu wezinkampani ezisebenzisa i-AI bezisebenza ngenkuthalo ukunciphisa izingozi ezifana nokungalungile ( Isimo se-AI ngo-2023: Unyaka wokuphuma kwe-Generative AI | McKinsey ) ( Isimo Se-AI: Inhlolovo Yomhlaba Wonke | McKinsey ), futhi lelo nani liyenyuka. Amafemu ezobuchwepheshe amise amabhodi ezimiso ze-AI; ohulumeni babhala imithethonqubo. Okubalulekile wukufaka izimiso zokuziphatha ekuthuthukisweni kwe-AI kusukela ekuqaleni (“Izimiso zokuziphatha ngokuklama”), kunokuba usabele kamuva.

Sengiphetha ngezinselele: ukunikeza i-AI ukuzimela okwengeziwe kuyinkemba esika nhlangothi zombili. Ingaveza ukusebenza kahle kanye nokuqanjwa okusha, kodwa idinga ibha eliphezulu lokuzibophezela. Iminyaka ezayo cishe izobona ingxubevange yezixazululo zobuchwepheshe (ukuthuthukisa ukuziphatha kwe-AI), izixazululo zezinqubo (inqubomgomo nezinhlaka zokwengamela), futhi mhlawumbe amazinga amasha noma izitifiketi (amasistimu e-AI angase ahlolwe futhi aqinisekiswe njengezinjini noma izinto zikagesi namuhla). Ukuzulazula ngempumelelo kulezi zinselele kuzocacisa ukuthi singahlanganisa ngokushelela kangakanani i-AI ezimele emphakathini ngendlela ekhulisa inhlalakahle nokwethenjwa kwabantu.

Isiphetho

I-Generative AI iguquke ngokushesha isuka ocwaningweni lwenoveli yaya kubuchwepheshe obuguqulayo benjongo evamile obuthinta wonke amagumbi ezimpilo zethu. Leli phepha elimhlophe lihlole ukuthi, ngo-2025, izinhlelo ze-AI sezivele zibhala kanjani izindatshana, ukuklama ihluzo, isoftware yokubhala amakhodi, ukuxoxa namakhasimende, ukufingqa amanothi ezokwelapha, ukufundisa abafundi, ukwenza kahle amaketanga okuhlinzeka, nokubhala imibiko yezezimali. Okubalulekile, kweminye yale misebenzi i-AI ingasebenza ngaphandle kokungenela komuntu , ikakhulukazi emisebenzini echazwe kahle, ephindaphindwayo. Izinkampani nabantu ngabanye sebeqala ukwethemba i-AI ukuthi yenze le misebenzi ngokuzimele, bathole izinzuzo ngesivinini nangesilinganiso.

Uma sibheka phambili ku-2035, simi onqenqemeni lwenkathi lapho i-AI izoba ngumhlanganyeli otholakala yonke indawo - ngokuvamile abasebenzi bedijithali abangabonakali abaphatha isimiso ukuze abantu bakwazi ukugxila kokuhlukile. Silindele ukuthi i-AI ekhiqizayo ishayele izimoto namaloli ngendlela ethembekile emigwaqeni yethu, ilawule i-inventory ezindaweni zokugcina izimpahla ngobusuku obubodwa, iphendule imibuzo yethu njengabasizi bomuntu siqu abanolwazi, inikeze iziyalezo zomuntu ngamunye kubafundi emhlabeni wonke, futhi isize nokuthola izindlela zokwelapha ezintsha kwezokwelapha - konke ngokugadwa okuqondile okunciphile. Umugqa phakathi kwethuluzi ne-ejenti uzofiphala njengoba i-AI isuka ekulandeleni iziyalezo ngokungenzi lutho ukuya ekukhiqizeni izixazululo.

Kodwa-ke, uhambo oluya kuleli kusasa le-AI elizimele kufanele lihanjiswe ngokucophelela. Njengoba sesichazile, isizinda ngasinye siletha isethi yaso yemikhawulo nezibopho:

• Ukuhlola Kwangempela Kwanamuhla: I-AI ayinaphutha. Iphumelela kakhulu ekuqashelweni kwephethini nasekukhiqizeni okuqukethwe kodwa ayinakho ukuqonda kwangempela nengqondo evamile ngomqondo womuntu. Ngakho, okwamanje, ukwengamela komuntu kuseyinetha lokuphepha. Ukuqaphela lapho i-AI ilungele ukundiza yodwa (futhi lapho ingekho khona) kubalulekile. Impumelelo eminingi namuhla iphuma kumodeli yeqembu labantu le-AI , futhi le ndlela yengxubevange izoqhubeka nokuba wusizo lapho ukuzimela okuphelele kungakabi khona ubuhlakani.

• Isithembiso Sakusasa: Ngokuthuthuka kwezakhiwo eziyimodeli, amasu okuqeqesha, nezindlela zokwengamela, amakhono e-AI azoqhubeka nokukhula. Iminyaka eyishumi elandelayo ye-R&D ingaxazulula amaphuzu amaningi obuhlungu amanje (ukunciphisa ukubona izinto ezingekho, ukuthuthukisa ukutolika, ukuqondanisa i-AI nezindinganiso zomuntu). Uma kunjalo, izinhlelo ze-AI ngo-2035 zingaba namandla ngokwanele ukuthi ziphathiswe ukuzimela okukhulu kakhulu. Ukuqagela kuleli phepha - kusukela kothisha be-AI kuya ikakhulukazi emabhizinisini azenzela wona - kungase kube iqiniso lethu, noma kudlulwe izinto ezintsha okunzima ukuzicabanga namuhla.

• Iqhaza Lomuntu Nokuzivumelanisa nezimo: Kunokuba i-AI ithathele abantu ngokuqondile, sibona kusengaphambili izindima eziguqukayo. Ochwepheshe kuyo yonke imikhakha bazodinga ukuthi babe nobuchule ekusebenzeni ne -AI - ukuyiqondisa, ukuyiqinisekisa, nokugxila ezintweni ezidinga amandla omuntu acacile njengozwela, ukucabanga kwamasu, nokuxazulula izinkinga eziyinkimbinkimbi. Imfundo nokuqeqeshwa kwabasebenzi kufanele kugxumeke ukuze kugcizelele lawa makhono abantu ahlukile, kanye nokufunda kwe-AI kwawo wonke umuntu. Abenzi bezinqubomgomo nabaholi bamabhizinisi kufanele bahlelele izinguquko emakethe yezabasebenzi futhi baqinisekise izinhlelo zokusekela zalabo abathintwe yi-automation.

• Izimiso Zokuziphatha Nokuphatha: Mhlawumbe okubaluleke kakhulu, uhlaka lokuziphatha kwe-AI nokubusa kufanele lusekele lokhu kukhula kobuchwepheshe. Ukwethemba uhlobo lwemali lokutholwa - abantu bazovumela i-AI ukuthi ishayele imoto noma basize ekuhlinzeni uma bethemba ukuthi iphephile. Ukwakha lokho kuthembana kuhilela ukuhlola okuqinile, ukubeka izinto obala, ukusebenzelana nababambiqhaza (isb, okubandakanya odokotela ekuklameni ama-AI ezokwelapha, othisha ngamathuluzi emfundo ye-AI), kanye nemithetho efanele. Ukusebenzisana kwamazwe ngamazwe kungase kudingeke ukuze kusingathwe izinselele ezifana ne-deepfakes noma i-AI empini, ukuqinisekisa ukuthi izinkambiso zomhlaba wonke zisetshenziswa ngendlela efanele.

Sengiphetha, i-AI ekhiqizayo imi njengenjini enamandla yenqubekelaphambili. Uma isetshenziswa ngobuhlakani, ingakhulula abantu ekubeni nesithukuthezi, ivule ubuhlakani, inikeze izinsiza eziqondene nawe, kanye nezikhala zamakheli (kulethe ubungcweti lapho ochwepheshe beyindlala khona). Okubalulekile wukusebenzisa ngendlela ekhulisa amandla omuntu kunokuwabeka eceleni . Ngokushesha, lokho kusho ukugcina abantu ku-loop ukuze baqondise i-AI. Ngokuhamba kwesikhathi, kusho ukufaka ikhodi kumagugu obuntu kumnyombo wezinhlelo ze-AI ukuze kuthi noma zisebenza ngokuzimela, zenze ngokwenzuzo yethu sonke.

Isizinda	Ukuzimela Okuthembekile Namuhla (2025)	Kulindelwe Ukuzilawula Okuthembekile ngo-2035
Ukubhala Nokuqukethwe	- Izindaba ezijwayelekile (ezemidlalo, umholo) zikhiqizwa ngokuzenzakalela.- Ukubuyekezwa komkhiqizo okufingqiwe yi-AI.- Okusalungiswa kwama-athikili noma ama-imeyili okuhlelwa kwabantu. ( UPhilana Patterson - ONA Iphrofayili Yomphakathi ) ( I-Amazon ithuthukisa ulwazi lokubuyekezwa kwamakhasimende nge-AI )	- Iningi lezindaba nokuqukethwe kokumaketha kubhalwe ngokuzenzakalela ngokunemba kweqiniso.- I-AI ikhiqiza ama-athikili aphelele nokukhishwa kwabezindaba okugadwa okuncane.- Okuqukethwe komuntu siqu okuphezulu okukhiqizwa ngokufunwa.
Ubuciko obubonakalayo nokuklama	- I-AI ikhiqiza izithombe kusuka ekwazisweni (umuntu ukhetha okungcono kakhulu).- Ubuciko bomqondo nokuhluka kwedizayini kudalwe ngokuzenzakalelayo.	- I-AI ikhiqiza izigcawu ezigcwele zevidiyo/zefilimu kanye nemifanekiso eyinkimbinkimbi.- Idizayini ekhiqizayo yemikhiqizo/imininingwane yomhlangano wezakhiwo.- Imidiya yomuntu siqu (izithombe, ividiyo) idalwe ngokufunwa.
Ukufakwa Kwekhodi Kwesoftware	- I-AI iqedela ngokuzenzakalela ikhodi futhi ibhala imisebenzi elula (ibuyekezwe yi-dev) - Ukwenziwa kokuhlola okuzenzakalelayo neziphakamiso zesiphazamisi. ( Ukubhala ngekhodi ku-Copilot: Idatha yango-2023 Iphakamisa Ingcindezi Eya Phansi Ngekhwalithi Yekhodi (okuhlanganisa nokubikezelwa kuka-2024) - I-GitClear ) ( I- GitHub Copilot Tops Report Report on AI Code Assistants -- Visual Studio Magazine )	- I-AI isebenzisa zonke izici ezicaciswe ngokuthembekile.- Ukulungisa iphutha okuzenzakalelayo nokugcinwa kwekhodi kumaphethini aziwayo.- Ukudalwa kohlelo lokusebenza lwekhodi ephansi ngokufaka okuncane komuntu.
Insizakalo yekhasimende	- Ama-Chatbots aphendula ama-FAQ, axazulule izinkinga ezilula (amacala ayinkimbinkimbi e-handoff).- I-AI isingatha ~70% wemibuzo evamile kwezinye iziteshi. ( Izibalo zesevisi yamakhasimende ezingama-59 ze-AI zika-2025 ) ( Ngo-2030, u-69% wezinqumo phakathi nokusebenzelana kwamakhasimende kuzoba ... )	- I-AI isingatha ukusebenzelana kwamakhasimende okuningi kusukela ekupheleni, okuhlanganisa imibuzo eyinkimbinkimbi.- Ukuthatha izinqumo ze-AI zesikhathi sangempela zemvume yesevisi (ukubuyiselwa kwemali, ukuthuthukiswa).- Ama-ejenti abantu kuphela ngokukhuphuka noma amacala akhethekile.
Ukunakekela impilo	- I-AI ibhala amanothi ezokwelapha; iphakamisa ukuxilongwa odokotela abaqinisekisayo.- I-AI ifunda ezinye izikena (i-radiology) ngokubheka; ihlola amacala alula. ( Imikhiqizo yokuthwebula yezokwelapha ye-AI Ingakhuphuka Kahlanu ngo-2035 )	- I-AI ixilonga ngokuthembekile izifo ezivamile futhi ihumusha izithombe eziningi zezokwelapha.- I-AI iqapha iziguli futhi iqale ukunakekela (isb., izikhumbuzi zemithi, izexwayiso eziphuthumayo).- "abahlengikazi" be-Virtual AI baphatha ukulandelwa okujwayelekile; odokotela bagxile ekunakekelweni okuyinkimbinkimbi.
Imfundo	- Abafundisi be-AI baphendula imibuzo yabafundi, bakhiqize izinkinga zokuzilolonga (abaqaphi bakathisha).- I-AI isiza ukugreda (ngokubuyekezwa kukathisha). ([I-Generative AI yemfundo ye-K-12	Umbiko Wocwaningo ka-Applify]( https://www.applify.co/research-report/gen-ai-for-k12#:~:text=AI%20tutors%3A%20Virtual%20AI,individual%20learning%20styles%20and%20paces ))
Logistics	- I-AI ithuthukisa imizila yokulethwa nokupakishwa (abantu babeka imigomo) - I-AI ihlaba umkhosi ubungozi bochungechunge lokuhlinzeka futhi iphakamise ukunciphisa. ( Izimo Eziphezulu Zokusebenzisa i-AI ku-Logistics )	- Ukudiliva okuzishayelayo kakhulu (amaloli, ama-drones) agadwe abalawuli be-AI.- I-AI ihlela kabusha ngokuzenzakalelayo ukuthunyelwa eduze nokuphazamiseka futhi ilungise uhlu lwempahla.- Ukuxhumana kochungechunge lokuhlinzekwa kokuphela ukuya ekupheleni (uku-oda, ukusatshalaliswa) okulawulwa yi-AI.
Ezezimali	- I-AI ikhiqiza imibiko yezezimali/izifinyezo zezindaba (ezibuyekezwe ngabantu) .- Abeluleki beRobo baphatha amaphothifoliyo alula; Ingxoxo ye-AI isingatha imibuzo yamakhasimende. ( I-Generative AI iza ngezezimali )	- Abahlaziyi be-AI bakhiqiza izincomo zokutshalwa kwezimali kanye nemibiko yengozi ngokunemba okuphezulu.- Ukuhweba okuzimele kanye nokulinganisa kabusha kwephothifoliyo ngaphakathi kwemikhawulo emisiwe.- I-AI ivuma ngokuzenzakalela ukubolekwa/izimangalo ezijwayelekile; abantu baphatha okuhlukile.

Izithenjwa:

1. Patterson, Philana. Izindaba zeholo ezizenzakalelayo ziyaphindaphindeka . I-Associated Press (2015) - Ichaza isizukulwane esizenzakalelayo se-AP sezinkulungwane zemibiko yemali engenayo ngaphandle kombhali ongumuntu ( Izindaba zeholo ezizenzakalelayo ziyaphindaphindeka | The Associated Press ).

2. McKinsey & Inkampani. Isimo se-AI ekuqaleni kuka-2024: I-Gen AI yokutholwa iyakhuphuka futhi iqala ukukhiqiza inani . (2024) - Ibika i-65% yezinhlangano ezisebenzisa i-AI ekhiqizayo njalo, cishe kabili kusukela ngo-2023 ( Isimo se-AI ekuqaleni kuka-2024 | McKinsey ), futhi idingida imizamo yokunciphisa ubungozi ( I-State of AI: Inhlolovo Yomhlaba Wonke | McKinsey ).

3. Gartner. Ngale Kwe-ChatGPT: Ikusasa Le-Generative AI Yamabhizinisi . (2023) - Ibikezela ukuthi ngo-2030, u-90% wefilimu ye-blockbuster ingase ikhiqizwe yi-AI ( I-Generative AI Use Cases for Industries and Enterprises ) futhi igqamisa izimo ezikhiqizayo zokusetshenziswa kwe-AI ezifana nomklamo wezidakamizwa ( I-Generative AI Use Cases for Industries and Enterprises ).

4. I-Twipe. Izindlela Eziyi-12 Izintatheli Ezisebenzisa Amathuluzi E-AI Egumbini Lezindaba . (2024) - Isibonelo se-"Klara" AI endaweni yezindaba ebhala u-11% wama-athikili, nabahleli abangabantu bebuyekeza konke okuqukethwe kwe-AI ( Izindlela Eziyi-12 zezintatheli Ezisebenzisa Amathuluzi e-AI Egumbini Lezindaba - Twipe ).

5. Izindaba ze-Amazon.com. I-Amazon ithuthukisa ulwazi lokubuyekezwa kwamakhasimende nge-AI . (2023) - Imemezela izifinyezo zokubuyekeza ezenziwe nge-AI emakhasini omkhiqizo ukusiza abathengi ( i-Amazon ithuthukisa ulwazi lokubuyekezwa kwamakhasimende nge-AI ).

6. Zendesk. 59 AI izibalo zesevisi yamakhasimende zango-2025 . (2023) - Ibonisa ngaphezu kwezingxenye ezimbili kwezintathu zezinhlangano ze-CX ezicabanga ukuthi i-AI ekhiqizayo izongeza "imfudumalo" enkonzweni ( izibalo zesevisi yamakhasimende ze-59 AI ze-2025 ) futhi ibikezela i-AI ku-100% yokusebenzisana kwamakhasimende ekugcineni ( izibalo ze-59 AI zesevisi yamakhasimende ze-2025 ).

7. I-Futurum Research & SAS. Umuzwa wango-2030: Ikusasa Lokuzizwisa Kwekhasimende . (2019) - Inhlolovo ethola ukuthi imikhiqizo ilindele ~ 69% yezinqumo phakathi nokubandakanyeka kwamakhasimende izokwenziwa imishini ehlakaniphile ngo-2030 ( Ukuze Ucabange Kabusha I-Shift ukuya ku-CX, Abakhangisi Kumele Benze Lezi Zinto Ezi-2 ).

8. Idathaiku. Izimo Eziphezulu Zokusebenzisa i-AI ku-Logistics . (2023) - Ichaza ukuthi i-GenAI ithuthukisa kanjani ukulayisha (yehlisa ~30% isikhala esingenalutho samaloli) ( I-Top Generative AI Use Cases in Logistics ) futhi ihlaba umkhosi ubungozi bochungechunge lokuhlinzeka ngokuskena izindaba.

9. I-Visual Studio Magazine. I-GitHub Copilot Iphakamisa Umbiko Wocwaningo Ngabasizi Bekhodi ye-AI . (2024) – Ukuqagela kokuhlela kwamasu kaGartner: ngo-2028, u-90% wabathuthukisi bebhizinisi bazosebenzisa abasizi bekhodi ye-AI (kusuka ku-14% ngo-2024) ( GitHub Copilot Tops Research Report on AI Code Assistants -- Visual Studio Magazine ).

10. Izindaba ze-Bloomberg. Sethula i-BloombergGPT . (2023) – Imininingwane yemodeli yepharamitha ye-Bloomberg engu-50B ehloselwe imisebenzi yezezimali, eyakhelwe kuTheminali ye-Q&A kanye nokusekelwa kokuhlaziya ( I-Generative AI iyeza ngezimali ).

Izindatshana ongathanda ukuzifunda ngemva kwalesi:

🔗 Imisebenzi I-AI Engakwazi Ukuyishintsha - Futhi Yimiphi Imisebenzi I-AI Ezoyishintsha?
Umbono womhlaba jikelele mayelana nesimo semisebenzi esiguqukayo, esihlola ukuthi yiziphi izindima eziphephile ekuphazanyisweni kwe-AI nokuthi yiziphi ezisengozini enkulu.

🔗 Ingabe i-AI Ingakwazi Ukubikezela Imakethe Yezitoko?
Ukungena okujulile emandleni, imikhawulo, kanye nokucatshangelwa kokuziphatha kokusebenzisa i-AI ekubikezelweni kwezimakethe zamasheya.

🔗 Ingasetshenziswa Kanjani I-Generative AI Ku-Cybersecurity?
Funda ukuthi i-AI yokukhiqiza isetshenziswa kanjani ukuze kuvikelwe izinsongo ze-inthanethi, kusukela ekutholweni okungaqondakali kuye ekumodeleni okusabisayo.