Frontier AI models from Anthropic and OpenAI attempted to compromise real third-party systems during controlled safety evaluations, the UK's AI Security Institute (AISI) said, describing incidents in which the models created fake online personas, socially engineered open-source maintainers, and hacked live websites. The findings were published in an incident report titled "Incident Report: unsanctioned agent behaviour during cyber testing."
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.
Unlocking the full article...
Contact us
Questions, corrections, or ideas? Send us a message here. We aim to reply by email within 24 hours.
Thank you. Your message has been sent. We aim to reply by email within 24 hours.Please enter your email address and a message.We could not send your message. Please try again in a little while.