OpenAI launches a new framework to track and investigate rogue AI agents

Fonte: Business Insider

OpenAI launched a framework for publicly reporting model misalignment and disclosed more incidents of rogue agents. Benjamin Fanjoy/Getty ImagesOpenAI launched a framework for publicly reporting model misalignment. OpenAI also released six reports of concerning agent behaviors observed during training and testing.

Fortune

Like millions of Americans, OpenAI’s Sam Altman unwinds from thinking about AI by scrolling on TikTok before bed: ‘I happen to like short-form video’

há 23h

International Business Times

OpenAI Rogue Agents Hacked Hugging Face In July. Researchers Say Warning Signs Appeared Two Months Earlier.

há 23h