Uzakwethu ophethe uhlelo olubalulekile lwezobuhlakani bokwenziwa kwaGoldman Sachs uxwayise ngokuthi ukusabalala ngokushesha kwe-AI eWall Street kunganciphisa ikhono lokucabanga lesizukulwane esilandelayo sabasebenzi bezezimali.
“Kukhona ingozi enkulu lapha: esikhathini se-AI, sidlulisela ikhono lokucabanga kulawa mamodeli, bese ekugcineni kuvela ukuncipha kwamakhono okuqonda, kuze kufike lapho nathi singasakwazi ukucabanga sisuka ezimisweni eziyisisekelo.” Kusho uChris Churchman, ophethe iMarquee, inkundla yedijithali yakwaGoldman Sachs yamakhasimende ezikhungo.
Ngokombhalo wenhlolokhono otholwe yi-CNBC, uChurchman usho la mazwi esiqeshini sakamuva se-podcast yakwaGoldman Sachs ethi “Exchanges”.
UChurchman uthe, njengoba ukuvela kobuchwepheshe besimanje kwenze abantu balahlekelwa kancane kancane amakhono okusebenzisa amabalazwe nokukhumbula, osomabhange nabo bangase balahlekelwe kancane kancane ikhono lokuhlaziya uma ama-algorithm ethatha wonke umsebenzi onzima wokuhlaziya.
“Ukucabanga kusabalulekile,” kusho yena. “Kusadingeka ucabange ngenkinga futhi uyihlele ibe impikiswano ephelele, kodwa manje sinikeza i-AI ikhono lokucabanga.”
UWall Street uqhubekisela phambili ukuhlanganisa i-AI ngokujulile kuzo zonke izingxenye zokuhweba nezamabhange, futhi lokhu kungaba “isivumelwano nodeveli”: esikhathini esifushane, i-AI ingakhuphula inzuzo yomkhakha wonke, kodwa esikhathini eside ingase icekele phansi isisekelo sabasebenzi abanamakhono umkhakha othembele kuwo ukuze uthuthuke.
Njengoba i-AI ithatha imisebenzi eminingi yansuku zonke eyayenziwa osomabhange nabahwebi abasebasha, laba basebenzi abasha babekwazi ukufunda ukucabanga nokwenza izinqumo ngale misebenzi. Manje, uma la mathuba okuqeqeshwa ethathwa yi-AI, izikhungo zezezimali zingase zidele isiko lomsebenzi eliguqula abasebenzi abasebasha babe ochwepheshe abanamava baseWall Street.
Lokhu kungase kwehlise nesidingo sezikhungo zezezimali sabasebenzi abasebasha kusukela ekuqaleni. I-CNBC yabika ngonyaka odlule ukuthi izikhungo zezezimali zaseWall Street zicwaninga ukuthi zingayisebenzisa kanjani i-AI ukwehlisa isilinganiso sabasebenzi abasebasha uma siqhathaniswa nabasebenzi abaphezulu.
UChurchman uthe amabhange kudingeka athole ibhalansi phakathi kokusebenzisa i-AI nokuvikela uhlelo lwendabuko lwaseWall Street lokufundisa “njengomeluleki nomfundi”. UChurchman wake waphatha ibhizinisi lokuhweba ngemali yangaphandle kwa-UBS, wajoyina iGoldman Sachs ngo-2021.
“Kudingeka ufunde ngokwenza, futhi ulwazi oluningi empeleni luyimfihlo, alukaze lubhalwe phansi.” kusho yena.
UChurchman uthe iGoldman Sachs kumele “iqinisekise ukuthi asilahlekelwa ulwazi olufihlekile nolwazi olusekelwe ekuzizweleni oluphethwe abanye babantu bethu abahamba phambili njengamanje, futhi ngesikhathi esifanayo iqinisekise ukuthi isizukulwane esilandelayo siyakwazi ukuthola la makhono”.
Ngokwesibonelo, umhwebi osaqala angasingatha izicelo zamakhasimende zamakhotheshini ngaphansi kokuqondisa komhwebi onamava, bese efunda ukwenza izinqumo zokuhweba ngale nqubo.
“Yebo singakwenza lokhu kube okuzenzakalelayo,” kusho yena, “kodwa-ke, sizokwazi yini ukuqeqesha abahwebi abaphezulu abaliqonda ngempela leli bhizinisi?”
UChurchman uthe uma kukhulunywa ngezinqumo ezinezingozi ezinkulu nezingaqinisekile kakhulu, uhlelo kumele luqinisekise ukuthi abasebenzi basenegunya lokugcina lokwenza izinqumo, esikhundleni sokubaguqula babe abantu abacindezela izinkinobho kuphela ohlelweni lwe-AI.
UChurchman uphinde abe ngusihlalo obambisene weqembu lomsebenzi le-AI kwaGoldman Sachs elibhekele ibhizinisi lamabhange nezimakethe emhlabeni wonke. Uthe ngisho neGoldman Sachs, engenye yamabhange otshalomali ahamba phambili emhlabeni, okwamanje ayikabi “nokuqonda okuphelele” kokuthi iluphathe kanjani lolu shintsho oseluqalile.
Kule nhlolokhono ye-podcast, uChurchman wabelane nangolwazi lweGoldman Sachs lokufaka i-AI enkundleni yeMarquee. IMarquee iqondiswe ezikhwameni ze-hedge nakwamanye amakhasimende ezikhungo, futhi isetshenziselwa ukufinyelela idatha yezimakethe yakwaGoldman Sachs, imibiko yocwaningo, ukuhlaziywa kwezingozi kanye nezinsizakalo zokwenza ukuhweba.
UChurchman uthe okwamanje inkundla ye-AI yeMarquee ivulekele kuphela abasebenzi bangaphakathi bakwaGoldman Sachs.
Ngokombono wezobuchwepheshe, inselelo enkulu ukuqinisekisa ukuthi izimpendulo ezinikezwa yi-AI zinembile ngo-100% futhi ziyakwazi ukuhlolwa. Uveze ukuthi ama-chatbot e-AI aqondiswe kubathengi ngokuvamile axwayisa abasebenzisi ngokuthi izimpendulo zingase zibe namaphutha, kodwa emkhakheni wezezimali osezingeni eliphezulu ukubekezelela amaphutha kuphansi kakhulu.
UChurchman uveze ukuthi ngesikhathi kwakhiwa inkundla ye-AI eqondiswe kumakhasimende, le software yaze yanikeza nokuzihlola okumangazayo.
“Lapho siqhubeka siyicindezela futhi siyihlola ngokuqinile, okungenani ithembekile.” kusho uChurchman. “Kunjengokungathi ithi: ‘Lalela, ekugcineni, ngingcono kakhulu ekwenzeni kubonakale sengathi ngiphelele kunokuba ngibe ophelele ngempela.’”