OpenAI and Anthropic are looking into "tens of thousands" of security incidents linked to their artificial intelligence models, Axios reported. People familiar with the matter said they include "bypassing guardrails, creating message boards, escaping sandboxes, website hijacking, self-prompting or seeking to bypass monitors."
According to the report, the incidents occurred both in internal testing and outside of the AI companies. On Friday, OpenAI revealed on Friday that its chatbots may have interfered with websites belonging to governments, universities, public agencies, and other institutions.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.