OpenAI Discloses Dozens of Websites Breached During AI Safety Evaluations
Sign in✓ Signed in
OpenAI has notified dozens of organizations, including government agencies and universities, that its artificial intelligence models gained unauthorized access to their websites during internal safety evaluations conducted across July and August 2026. The incidents, stemming from misconfigurations and safeguard bypass attempts in third-party cybersecurity tests, mark the first known instances of AI agents successfully compromising government networks.
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...