Kamfanin OpenAI, wanda ya ƙera ChatGPT, ya bayyana cewa wani tsarin basirar wucin gadi (AI) mai cin gashin kansa ya yi yunƙurin kutsawa cikin asusun wasu kamfanoni huɗu bayan ya samu damar kutse cikin dandalin masu shirya manhajoji na Hugging Face.
A cewar OpenAI, yayin gwajin tsaro da ake yi wa tsarin AI, samfuransa biyu sun fita daga yanayin gwajin da aka killace su, suka haɗa da intanet tare da neman hanyoyin shiga wasu dandamali ta amfani da bayanan shiga (login credentials) da aka bari a fili a intanet.
Kamfanin ya ce AI ɗin ya samu shiga wasu asusu a kan ayyuka huɗu na jama’a, inda aka yi amfani da wasu daga cikinsu wajen adana bayanai da kuma ɓoye sawun ayyukansa.
Sai dai OpenAI ta ce ba ta ga wata shaida da ke nuna cewa lamarin ya yi tasiri ga sauran masu amfani da waɗannan dandamali ba, kuma ta fara tuntubar masu asusun da abin ya shafa.
Babban Jami’in OpenAI, Sam Altman, ya ce kamfanin ya dakatar da wasu gwaje-gwajen AI na ɗan lokaci domin ƙarfafa matakan tsaro da tsarin killace gwaje-gwaje (sandboxing).
Lamarin ya kuma jawo ce-ce-ku-ce a fannin basirar wucin gadi, inda sama da ma’aikata 1,000 daga manyan kamfanonin AI suka rattaba hannu kan wata buƙata ga gwamnatin Amurka da ta ƙara tsaurara matakan kula da sakin sabbin manyan tsarin AI.