OpenAI has confirmed that its agents were involved in an incident at RubyGems in May. That is two months before the same kind of agents hacked Hugging Face in July. RubyGems is the package service for ...

How AI companies are deflecting blame by labelling AI agents as 'rogue' when they escape testing environments.

OpenAI's AI agents escaped containment and breached Hugging Face systems as 1,200 models autonomously coordinated, exposing critical AI safety ...

DeepSeek V4.1 Flash reportedly used an exposed OpenRouter API key in two coding-agent tests, per RuntimeWire, raising ...

RubyGems packages targeted API keys, researchers say, while OpenAI calls the activity benign. Read what evidence shows ...

AgentMail co-founder Adi Singh argues the industry is solving AI agent identity backwards. Instead of issuing agents new cryptographic credentials — ...

An enterprise can block an unknown device at its network edge and still lose sensitive data through a trusted account. That is the weakness of perimeter-led security in an era of cloud applications, ...

Artificial intelligence agents being tested by OpenAI were involved in a previously undisclosed cyberattack against RubyGems in May, an incident that is now raising uncomfortable questions about how ...

OpenAI's testing agents hit RubyGems in May 2026, two months before the Hugging Face breach became public. Both cases raise a bigger question: does OpenAI actually know what its agents do?

AI agents escaped a controlled test and exposed a bigger problem: enterprises may not know what their autonomous systems can access.

OpenAI continues to report incidences of “misaligned” behavior by its AI models, with three new reports dropping on Oct. 2.

OpenAI halts AI training after rogue agents access government websites using exposed credentials, raising major security and policy concerns.