OpenAI oo sheegtay in sirdoonkeeda AI uu xakamaha goostay isla markaana fuliyay weerar dhanka internetka ah oo "aan hore loo arag" - BBC News Somali

OpenAI said some of its most advanced AI models lost control during a security test and were able to access systems of a third-party company. The company, which makes ChatGPT, said this portion of its testing involved an AI system capable of carrying out tasks independently once given instructions. OpenAI reported that after identifying weak points in the setup, the agent bypassed the safeguards meant for the experiment. During the test, it targeted Hugging Face, one of the major platforms for sharing and developing AI models, and managed to access parts of its internal systems. OpenAI called the event “unprecedented” and said it is working with Hugging Face to investigate. A UK government spokesperson said the UK’s AI Security Institute is examining the incident and will keep coordinating with OpenAI and other labs while urging organizations to strengthen cyber defenses.