I-Florida imangalele i-OpenAI kanye noSam Altman ngokubeka izingane nezintsha engcupheni enkulu, ngokungasebenzisi izindlela eziphumelelayo zokuqinisekisa iminyaka. Le ncwadi yomthetho, eyafakwa ngoMsombuluko (1 Juni), ithi i-ChatGPT ikhuthaza ukuncika kubasebenzisi abancane futhi ikhuthaze izenzo eziyingozi. Ummeli jikelele uJames Uthmeier wathi ngesikhathi sezindaba inkampani ikhohlise abazali ngokuveza uhlelo njengokuphepha. I-OpenAI ayikakasho lutho ngokusemthethweni ngalawa macala.
Ukungabikho kwezindlela zokuqinisekisa iminyaka
Ngokwencwadi yomthetho, inguqulo yamahhala ye-ChatGPT ayinalo uhlelo lokulawula noma lokuqinisekisa iminyaka, okuvumela ngisho izingane ezingakafinyeleli eminyakeni eyi-13 ukuthi zisebenzise le nkonzo ngaphandle kwemikhawulo. Enguqulweni ekhokhelwayo, nakuba kunenkambu yokufaka iminyaka, azikho izindlela zokuqinisekisa le mininingwane noma zokwazisa abazali mayelana nokuxhumana okwenziwa izingane. Ummeli jikelele uJames Uthmeier wagcizelela ukuthi inkampani yayazi ukuthi izingane zisebenzisa i-ChatGPT, kodwa ayizange ithathe izinyathelo zokuvimbela ukufinyelela. Leli cala elisemthethweni likhomba lokhu ukungayinaki njengokunganakwa okuhleliwe kokuphepha kwezingane.
Ucwaningo lwaseyunivesithi lweseka izinsolo zomonakalo
Ukuqinisa lezi zinsolo, abashushisi bacaphuna ucwaningo lwakamuva lwe-Drexel University e-United States, olwahlola umthelela we-chatbots entsheni. Ucwaningo lwahlola abasebenzisi benkundla yomncintiswano, i-Character.AI, futhi lwathola imiphumela efana nokulahlekelwa ubuthongo, ukwehla ekusebenzeni esikoleni, nokuncipha kokuxhumana nomphakathi. Nakuba ukugxila kwaleli cala kuse-ChatGPT, umnyango womthetho uthi amaphethini afanayo okuncika ayasebenza kumsizi we-OpenAI. Incwadi yomthetho igcizelela ukuthi ubuhlakani bokwenziwa bulingisa uzwelo nezici zomuntu ukuze kukhohlise abasebenzisi futhi kutholwe ulwazi oluningi oluyimfihlo.
Izinkombandlela eziyingozi nezinsongo empilweni yengqondo
Enye into ebalulekile yaleli cala umbiko we-Center for Countering Digital Hate (CCDH), owalingisa izingxoxo ne-ChatGPT ngokwenza sengathi umuntu oyingane. Kulezi zingxoxo, i-chatbot inikeze izeluleko ezinemininingwane mayelana nokufihla ukudla, kanye nezinyathelo zokuhlela ukuzibulala nokuzilimaza. Umbiko we-CCDH wembula ukuthi umsizi akazange aphendule imibuzo ebucayi kuphela, kodwa wanikeza nezinyathelo ezithile zokwenza lezo zenzo. Umnyango womthetho ubheka lezi zingxoxo njengobufakazi obuqondile bengozi lemvelaphi kubantu abancane abasengozini.
Izinyathelo zakamuva ze-OpenAI zithathwa njengezinganele
NgoJanuwari, i-OpenAI yethula uhlelo olulinganisa iminyaka yabasebenzisi ngokusekelwe ezindleleni zokuziphatha, futhi lusebenzisa izinyathelo ezengeziwe zokuvikela lapho kutholakala ingane. Njengamanje, imithetho yenkampani ivimbela ukusetshenziswa kwe-ChatGPT yizingane ezingaphansi kweminyaka eyi-13 futhi idinga imvume yabazali entsheni ephakathi kweminyaka eyi-13 neyi-17. Kodwa-ke, ummeli u-Uthmeier uthi lezi zinyathelo zifike sekwephuzile futhi azikaqinisekisi ukuqinisekisa okuqinile. Icala lisiza ukuthi ukungabikho kokugadwa okusebenzayo kuvumela izingane ukuthi zigweme le mikhawulo kalula.
Ummeli jikelele waseFlorida wathi i-OpenAI, i-ChatGPT, noSam Altman mathupha bangase babhekane nomonakalo ongafinyelela ezigidini zezigidi zamadola. “Sikholelwa ukuthi i-OpenAI, i-ChatGPT yayo, kanye noSam Altman mathupha, banomthwalo wemfanelo ngenani elingafinyelela ezigidini zezigidi zamadola,” kusho u-Uthmeier ngesikhathi sezindaba. Leli cala elisemthethweni lifuna ukubophezela inkampani hhayi kuphela ngomonakalo osuwenziwe, kodwa nangokuphoqa ukuthi isebenzise izindlela eziqine kakhulu zokuqinisekisa. Leli cala likhulisa impikiswano mayelana nomthwalo wemfanelo wezinkampani zobuhlakani bokwenziwa ngomthelela wemikhiqizo yazo emiphakathini esengozini.