Ucwaningo Olusha LweHarvard Lubonisa I-AI Ingase Ithathe Isikhundla Abaphathi Bezikhwama Eziningi

I-Algorithm Ehhovisi Lekhona: I-AI Icabanga Ukudlula Abaphathi Besikhwama Sabantu

Emashumini eminyaka, imboni ye-mutual fund ithengise isithembiso esihehayo: nikeza imali yakho kumhlaziyi ongumuntu okhaliphile, umuntu ochithe iminyaka engu-20 efunda amashidi ebhalansi, ehlezi phezu kwezingcingo zeholo, futhi ethuthukisa umuzwa cishe onembile mayelana nokuguquguquka kwemakethe - futhi bazokwenza ngcono imakethe. Leso sithembiso besilokhu sintekenteke. Manje, ucwaningo oluyingqopha-mlando oluphuma e-Harvard Business School lusongela ukusihlakaza ngokuphelele. Abacwaningi bathole ukuthi ubuhlakani bokwenziwa bungakwazi ukubikezelau-71% wohwebo lwe-mutual fundngokunemba okuphawulekayo, kuphakamisa umbuzo obungabonakala kungenangqondo eminyakeni emihlanu edlule: uma umshini ungalindela ukuthi umphathi wesikhwama uzokwenzani ngaphambi kokuba akwenze, yini ngempela abatshalizimali abayikhokhelayo?

Izingqinamba ziyaguquguquka ngale kwe-Wall Street. Lena indaba mayelana nokuthi kwenzekani lapho ukuqashelwa kwephethini - ikhono eliwumgogodla lanoma yimuphi uchwepheshe - kuba yimpahla. Futhi yindaba wonke umholi webhizinisi, hhayi nje ochwepheshe bezezimali, okudingeka bayiqonde njengamanje.

Lokho Okutholwe Empeleni Ucwaningo Lwe-Harvard

Ucwaningo lwe-Harvard luqeqeshe amamodeli okufunda emishini eminyakeni yomlando wedatha yokuhweba, ukudalulwa kwezimali, namasignali emakethe. Amamodeli ayengagcini nje ngokuhlonza izitayela zomkhakha obanzi; babebikezela izinqumo ezithile zephothifoliyo zabaphathi bezikhwama ngabanye - yiziphi izitoko ababezozithenga, bazozinquma nini, futhi nini. Izinga lokunemba lokubikezela elingu-71% esizindeni esiyinkimbinkimbi nesinomsindo njengokuphatha iphothifoliyo esebenzayo liyamangalisa. Ngokomongo, imodeli ebikezela ukuphenduka kohlamvu lwemali kungaba okulungile ngo-50% wesikhathi ngenhlanhla yodwa.

Okwenza okutholakele kucace kakhulu ukuthi kuveza obala izinto ezisetshenziswa abaphathi bezikhwama ezikhokhelwa kakhulu. Kunokuba kusetshenziswe ukuqonda okuyinoveli kwangempela, ingxenye enkulu yabaphathi abasebenzayo ibonakala iwukuziphatha okuqhutshwa iphethini - ukuphendula ezimangalisweni ezifanayo zemiholo, izimpawu ezifanayo zomfutho, izinkomba ezifanayo ezinkulu ngezindlela ezingabikezelwa. I-AI yayingadingi ukuqonda kungani umphathi angenza ukuhweba. Kwamane kwafunda ukuqaphela izimo ababehlala ngaphansi kwazo ngokuthembekile.

Lokhu kuhambisana nocwaningo lwangaphambilini. Umbiko wango-2022 we-S&P Dow Jones Indices uthole ukuthi esikhathini esiyiminyaka engu-20, ngaphezulu kuka-94% abaphathi bezikhwama ezinkulu zase-U.S. abasebenzayo benza ngaphansi kokusebenza kahle kwenkomba yabo yebhentshimakhi. Okutholwe yi-Harvard kwengeza ungqimba olusha: akukhona nje ukuthi abaphathi abaningi abasebenzayo abahluleki nje ukushaya imakethe, izinqumo zabo zingase zibe ngomshini ngokwanele ukuze i-algorithm ilingise — ngengxenyana yezindleko.

Kungani Ukubikezela Okungu-71% Kuyinkinga Yebhizinisi, Hhayi Inkinga Yezezimali Kuphela

Ochwepheshe bezezimali bangalingeka ukuthi baphathe lokhu njengenhlekelele eqondene nemboni ethile. Bayoba iphutha. Ucwaningo lwe-Harvard luyiphoyinti ledatha ngephethini enkulu kakhulu: Amasistimu e-AI aya ngokuya ekwazi ukuphindaphinda ukwahlulela kochwepheshe kunoma yisiphi isizinda lapho izinqumo zilandela imithetho efundekayo, nanoma leyo mithetho ingabhalwanga ngokusobala noma kuphi.

Cabangela ukuthi ukuphathwa kwesikhwama okusebenzayo kanye nokuphathwa kwebhizinisi okujwayelekile kufana ngani. Kokubili kuhilela ukuqoqa ulwazi, ukuhlonza amaphethini, ukusebenzisa ama-heuristics akhiwe isipiliyoni, kanye nokwenza izinqumo ngaphansi kokungaqiniseki. Uma i-AI ikwazi ukwenza imodeli yenqubo yokuthatha izinqumo zomphathi wesikhwama ngokunemba okungama-71%, ingabonisa ngokusobala ingxenye enkulu yezinqumo ezenziwa abaphathi bemisebenzi, abaqondisi bakwa-HR, abaholi bezentengiselwano, nabahlaziyi bebhizinisi — abantu abanobuhlakani babo obusekelwe ekuboneni nasekuphenduleni amaphethini.

"Usongo kubasebenzi bolwazi akukhona ukuthi i-AI izothatha indawo yokwahlulela komuntu ngokuphelele - ukuthi i-AI izothatha indawo yezingxenye zokwahlulela komuntu empeleni ezifanisa iphethini nje. Futhi lokho kuvele kube ingxenye enkulu ngokumangalisayo."

Lokhu akusho ukuthi ubuchwepheshe bomuntu buba yize. Kusho ukuthi imvelo yobuchwepheshe obubalulekile iyashintsha. Abaphathi bezikhwama abayosinda futhi baphumelele yilabo abenza okuthile i-AI engenakukwazi ukuyiphindaphinda kalula: hlanganisa ulwazi olusha lwangempela, akhe ubudlelwano obudala izinzuzo zolwazi, futhi asebenzise ukwahlulela ezimeni ukuze angabi naso isibonelo esingokomlando. Umqondo ofanayo uyasebenza kuzo zonke izizinda zobungcweti manje esezibunjwa kabusha ubuhlakani bomshini.

Ukuphazamiseka Kwe-AI Kwezimboni Eziseduze Kakhulu

Imboni ye-mutual fund empeleni iyi-canary emayini yamalahle ye-white-collar automation. Inothile ngedatha, inamamethrikhi okusebenza acacile, futhi ibingaphansi kwengcindezi yezindleko ezivela ezikhwameni zenkomba ezingenzi lutho iminyaka - okuyenza yamukele ngokungavamile ukwamukelwa kwe-AI. Ezinye izimboni zibhekisisa kahle.

Ekunakekelweni kwezempilo, amasistimu e-AI okuxilonga afana ne-Google's DeepMind abonise amandla okuthola izifo ezithile zamehlo nomdlavuza ngokufana okunembile noma okudlula odokotela abangochwepheshe. Ngokomthetho, amathuluzi akhelwe kumamodeli amakhulu ezilimi enza imisebenzi yokubuyekeza inkontileka ebidinga abantu abasebenzisana nabo abancane abasebenza ubusuku bonke. Ekubalweni kwezimali nasekuhleleni izimali, izinkundla eziqhutshwa yi-AI ziwukuhlaziya okuzenzakalelayo kokuhlukahluka, ukubikezela kokugeleza kwemali, kanye nokutholwa okudidayo okwake kwadinga isikhathi somhlaziyi ophezulu.

Uchungechunge olujwayelekile akukhona ukuthi i-AI ihlakaniphe kunochwepheshe kule mikhakha. Ukuthi i-AI ayikhathali, ayiguquki, futhi ishibhile kakhulu ukukala. Umphathi wesikhwama somuntu angase abize inkampani engu-$500,000 ngonyaka ngomholo, izinzuzo, kanye nemali engaphezulu. Uhlelo lwe-AI olukwazi ukubikezela u-71% wokuhweba kwalowo mphathi lusebenza ngengxenye yalezo zindleko - futhi aludingi ibhonasi, uhlelo lwesabatha, noma lokulandelana.

Yini Esindayo Kuhlelo Lokusebenza: Incazelo Entsha Yevelu Lomuntu

Ukusabela okungokwemvelo ocwaningweni olufana nalokhu kuyavikela: ukuphikisa ngokuthi ukwahlulela komuntu akunakuthathelwa indawo, ukuthi i-AI ayikwazi ngempela ukuqonda umongo, ukuthi kuyohlale kunendima yochwepheshe abanolwazi. Okunye kwalokho kuyiqiniso. Kodwa impendulo ekhiqiza kakhulu iwukucacisa kahle ukuthi yiziphi izici zobuchwepheshe bomuntu ezihlala zinzima ngempela ukuzenza.

Ngokusekelwe kumzila wamanje wekhono le-AI, lawa makhono alandelayo abonakala eqinile:

• Ukuthembeka okusekelwe ebudlelwaneni: Amaklayenti nababambiqhaza bavame ukwenza izinqumo ngokusekelwe kulokho abamethembayo, hhayi nje ukuthi yiluphi ulwazi abalutholayo. Ukwethembana kwakhelwa ngokusebenzisana kwabantu okuqhubekayo kanye nokuqondanisa okubonisiwe kwezithakazelo — hhayi umphumela we-algorithmic.
• Isahlulelo sezimiso zokuziphatha nesokulawula: Izimo zokuzulazula lapho imithetho ingacacile, ukungqubuzana kwezithakazelo zabathintekayo, noma izimo zenoveli zidinga ukucabanga ngokuziphatha kusafuna ukuziphendulela komuntu.
• Ukwakhiwa kobuciko: Ukuhlanganisa imininingwane evela ezizindeni ezihlukene - ukubona ukuthi ithrendi ekuziphatheni kwabathengi ixhumeka ekubeni sengozini ye-supply chain ixhuma kumthethonqubo osavelayo - kudinga uhlobo lokucabanga lokuhlanganisa i-AI isingatha ngokuthembekile kunokuqashelwa kwephethini.
• Ukuxhumana nababambiqhaza: Ukuhumushela ukuhlaziya okuyinkimbinkimbi kube yizindaba ezixoxwayo ezigqugquzela isenzo - ukukholisa ibhodi, ukuthulisa iklayenti elikhathazekile, ukukhuthaza iqembu - kuyinselelo enkulu yokuxhumana nabantu.
• Ukuphatha izinto ezintsha zangempela: Uma izimo ziphakama ngaphandle kwesibonelo esingokomlando (ubhubhane lomhlaba wonke, ukushaqeka kwezwe, ubuchwepheshe obushintshashintshayo), ukuzivumelanisa nezimo nobuciko bomuntu kubaluleke kakhulu kunokuba kungezelelwe.

Abaphathi bezikhwama asebevele bajwayelana nalokhu abazami ukuncintisana nama-algorithms ngesivinini sokukhethwa kwesitoko noma umthamo wokucubungula idatha. Bazibeka njengabadwebi bephothifoliyo, abaphathi bobudlelwano bamakhasimende, nabaphathi bezinhlaka eziyinkimbinkimbi zezingozi - izindima ezidinga ukuba khona komuntu nokuziphendulela, hhayi nje ikhono lokufanisa iphethini.

Indlela Izinhlangano Ezibheke Phambili Eziphendula Ngayo

Impendulo ehlakaniphe kakhulu yokuphazamiseka kwe-AI ayikona ukuphika noma ukwethuka — iwukuhlanganiswa. Izinhlangano ezizosebenza kahle kakhulu kule minyaka eyishumi ezayo yilezo ezisebenzisa i-AI ukuze zisuse umsebenzi wokufanisa iphethini onenani eliphansi kuyilapho zisatshalaliswa kabusha ithalente lomuntu emisebenzini ehlala inzima ngempela ukuzenza ngokuzenzakalelayo.

Eqinisweni, lokhu kusho ukwakha ingqalasizinda yokusebenza enikeza abantu ukufinyelela kubuhlakani obukhiqizwa yi-AI ngaphandle kokubadinga ukuthi babe ososayensi bedatha ngokwabo. Umholi wezentengiso kufanele akwazi ukubona amaphuzu okuhola aqhutshwa yi-AI eduze nomsebenzi we-CRM ngaphandle kokuguqula phakathi kwamapulatifomu amahlanu ahlukene. Umqondisi wakwa-HR kufanele akwazi ukuveza amasignali engcuphe yokugcinwa evela kudatha yabasebenzi ngaphandle kokwakha mathupha amadeshibhodi. Umsebenzisi wezezimali kufanele akwazi ukwenza izibikezelo zesimo sezulu ngokuhamba kwemali ngaphandle kwethimba lomhlaziyi elizinikele.

Lena kanye ifilosofi engemuva kwenkundla efana ne-Mewayz, ehlanganisa amamojula okuphatha amabhizinisi angaphezu kuka-200 - ahlanganisa i-CRM, ama-invoyisi, i-HR, i-payroll, analytics, ukuphathwa kwemikhumbi, nokunye okwengeziwe - endaweni eyodwa yokusebenza. Uma imininingwane eqhutshwa yi-AI ikhona ngaphakathi kwenkundla efanayo lapho izinqumo zithathwa khona, kunokuba zithuliswe ngethuluzi elihlukile, iluphu yempendulo phakathi kobuhlakani nesenzo iqina kakhulu. Emabhizinisini angu-138,000 asebenzisa i-Mewayz emhlabeni wonke, lokho kuhlanganiswa akusona isifiso sesikhathi esizayo; kuyiqiniso elisebenzayo lamanje.

Izindleko Zokulinda: Ukuthi Ukungasebenzi Okubukeka Kanjani Eminyakeni emihlanu

Kunokuthambekela ezimbonini eseziqalile ukuphatha ukuphazamiseka kwe-AI njengegagasi elihamba kancane — into okufanele iqashwe endaweni enethezekile kuyilapho uqhubeka nebhizinisi njengenjwayelo. Ucwaningo lokuphathwa kwesikhwama seHarvard luyisikhumbuzo sokuthi amagagasi angahamba ngokushesha kunalokho abakulindele. Imboni ye-mutual fund yachitha iminyaka ichitha izimali zenkomba ezingenzi lutho njengomkhiqizo we-niche wabatshalizimali abangahlakaniphile. Ngo-2023, i-passive funds yase idlule izimali ezisebenzayo eziphelele zezimpahla ezingaphansi kokuphathwa e-United States ngokokuqala ngqa emlandweni.

Amabhizinisi nezingcweti ezisengozini enkulu yokuphazamiseka kwe-AI akubona labo abasemikhakheni yobuchwepheshe ngokusobala - yilabo abazakhele indawo yabo yokuncintisana ekufinyeleleni okukhethekile olwazini noma ekhonweni lokucubungula nokuhumusha idatha ngokushesha kunabaqhudelana nabo. Zombili lezo zinzuzo ziguguleka ngokushesha lapho i-AI ingena esithombeni. Inzuzo yolwazi olukhethekile iyanyamalala lapho i-AI ingahlanganisa idatha yomphakathi ngezinga. Inzuzo yokucubungula iyanyamalala uma i-AI ikwazi ukuqalisa ukuhlaziya ngemizuzwana ethathe amaviki ngaphambilini.

Okungonakalisi - futhi empeleni kuba yigugu kakhulu - yikhono lokubuza imibuzo engcono, ukwakha ubudlelwano bangempela, kanye nokusebenza ngaphakathi kwamasistimu ahlanganisiwe ahumusha ukuqonda kube ukwenza ngaphandle kokungqubuzana. Izinhlangano ezitshala imali kulolo hlobo lwengqalasizinda namuhla azilungiseleli nje ukuphazamiseka kwe-AI. Bakha imodeli yokusebenza ezochaza ukusebenza kwebhizinisi kwesizukulwane esilandelayo.

Isifundo Sangempela Kukubalwa kwe-AI ye-Wall Street

Ucwaningo lwase-Harvard luzokhipha izihloko zezindaba mayelana namarobhothi athatha isikhundla sabaphathi bezikhwama, futhi lezo zihloko ngokuvamile zizogeja iphuzu. Okutholakele okubaluleke kakhulu akukhona ukuthi i-AI ingakwazi ukuphindaphinda izinqumo zochwepheshe - ukuthi into ebiza kakhulu ngezinqumo zochwepheshe kuphenduke izingxenye umshini ongakwazi ukuzisingatha ngokushibhile. Lokho kuqaphela kushintsha umnotho wobuchwepheshe kuyo yonke imboni, hhayi nje ezezimali.

Ochwepheshe nezinhlangano ezizophumelela yilabo abamukela leli qiniso ngaphandle kokukhubazwa yilo. Bazoklama kabusha izindima zabo ezizungeze izici zomuntu zangempela - ukwethembana, ukusungula izinto, ukwahlulela okuhle, ubuhlakani bobudlelwano - kuyilapho bamukela i-AI njengenjini ephethe ukuqashelwa kwephethini, ukuhlanganiswa kwedatha, nokubikezela okujwayelekile. Bazotshala imali ezinkundleni zokusebenza ezididiyelwe ezenza ubuhlakani obukhiqizwe nge-AI busebenze ngokushesha, kunokuba bubuthathe njengesengezo ekugelezeni komsebenzi okukhona.

Abaphathi bezikhwama ezihlanganyelwe abasinda eshumini leminyaka elizayo ngeke kube yilabo abangayinaki i-algorithm. Bazoba yilabo abafunda ukusebenza eceleni kwayo - besebenzisa i-AI ukuphatha i-71% ebikezelwayo ukuze bakwazi ukugxila ngokuphelele ku-29% engalindelekile lapho ukwahlulela komuntu kusawenza wonke umehluko. Leyo zibalo iyasebenza kuwo wonke umholi webhizinisi ozulazula kuguquko lwe-AI njengamanje. Umbuzo awukona ukuthi uzivumelanise yini. Umbuzo uthi ungaqala ngokushesha kangakanani.

Imibuzo Evame Ukubuzwa

Ingabe i-AI ingakwazi ngempela ukubikezela ukuhweba kwe-mutual fund kangcono kunabaphathi abangabantu abanolwazi?

Ngokocwaningo lwe-Harvard Business School, amamodeli e-AI angabikezela cishe u-71% wohwebo lwe-mutual fund ngokunemba okumangalisayo. Lezi zinhlelo zihlaziya amasethi edatha amaningi - amaphepha ebhalansi, izingcingo zeholo, amasiginali wezomnotho omkhulu - ngokushesha kakhulu kunanoma yimuphi umhlaziyi ongumuntu. Nakuba lokho akuqinisekisi imbuyiselo ephakeme kuzo zonke izimo zemakethe, kuphakamisa ngokuqinile ukuthi i-AI inomkhawulo olinganisekayo, wesakhiwo ngaphezu kokuphathwa kwesikhwama esivamile ekuqashelweni kwephethini nokuvumelana kwezinqumo.

Kusho ukuthini lokhu kubatshalizimali bansuku zonke abafaka imali ezikhwameni eziphathwayo?

Kuphakamisa imibuzo ebucayi mayelana nokuthi izinkokhelo zeprimiyamu ezikhokhiswa abaphathi bezikhwama abasebenzayo zifanelekile yini. Uma i-AI ingakwazi ukuphindaphinda futhi ibe namandla okudlula amasu abo, abatshalizimali bangase banikezwe kangcono izimoto eziqhutshwa i-algorithm noma ezingenzi lutho. Lolu shintsho luphinde lugcizelele ukubaluleka kokusebenzisa ibhizinisi elihlakaniphile namathuluzi ezezimali ukuze uphathe imali yakho ngendlela ephumelelayo, kunokuthembela ngokuphelele kubaxhumanisi abangabantu ababukhali babo buba buncane.

Abanikazi bamabhizinisi amancane nosomabhizinisi bangayisebenzisa kanjani i-AI ukuze benze izinqumo ezihlakaniphile zezimali?

Amapulatifomu afana ne-Mewayz — isistimu yokusebenza yebhizinisi enamamojula angu-207 etholakala ku-app.mewayz.com ngo-$19 kuphela/ngenyanga — inika osomabhizinisi ukufinyelela kumathuluzi anamandla e-AI ake atholakala emabhizinisini amakhulu kuphela. Kunokuba bakhiphe ukwahlulela kwezezimali kubeluleki ababizayo, abanikazi bamabhizinisi bangasebenzisa izibalo ezihlanganisiwe ukuze baqaphe ukuhamba kwemali, imodeli yezimo, futhi benze izinqumo ezisekelwe idatha ngokuqina okufanayo okuhlelekile manje okuphazamisa imboni yokuphatha isikhwama se-Wall Street.

Ingabe kukhona ukulinganiselwa kulokho i-AI engayenza njengamanje ezimakethe zezimali?

Yebo. I-AI iphumelela kakhulu ekuhlonzeni amaphethini omlando nasekucubunguleni idatha ehleliwe, kodwa ingazabalaza nemicimbi yedada elimnyama engakaze ibonwe, ukushaqeka kwezwe, noma amashifu aqhutshwa isayense yomuntu ewela ngaphandle kwedatha yayo yokuqeqeshwa. Abaphathi abangabantu basaletha ukwahlulela ngokomongo, ukucabanga ngokuziphatha, kanye nokucabanga okuguquguqukayo phakathi nokugudluzwa kwezimakethe ngokwedlulele. Umphumela ongase ube seduzane kakhulu imodeli eyingxube, lapho i-AI isingatha khona ukuhlaziya kuyilapho abantu begcina ukwengamela izinqumo ezibaluleke kakhulu.