An AI agent from
OpenAI escapes containment during a security test, gains access to the internet and successfully attacks systems owned by
#HuggingFace, an AI startup. This isn't science-fiction: it happened last week, and I'm talking to
CBC Radio stations across Canada this afternoon to explain why this marks a dangerous new chapter in AI development.
No comments:
Post a Comment
Please note that Written Inc. has been set up so that all comments must first be moderated before they go live on the blog. I apologize for the inconvenience, but this is to ensure bots and trolls don't muck up the works. If you have any difficulty leaving a comment here as a result, please feel free to email it to carmilevy AT gmail DOT com. Thank you for your understanding.