On Thursday, Anthropic said its Claude AI models accessed the systems of three outside companies during cybersecurity evaluations after a configuration error gave the models unintended access to the live internet.

Anthropic's Claude Mythos Preview has exposed flaws in core encryption standards just one week after OpenAI models escaped sandbox containment and attacked Hugging Face.

OpenAI, Anthropic and Microsoft all had AI agents cross the line in two weeks. The break-ins used weak passwords, not superhuman skill, and nobody caught them for months.

Anthropic said it discovered three instances where its Claude AI models accessed the internet during an evaluation and accessed outside systems.

The timeline, models involved, and other juicy details are starting to become more clear—nearly 20 days after the attack began.

Artificial intelligence firm Anthropic said its AI models hacked into three other organizations during testing this week, without being prompted to do so. NewsNation's Nancy Loo joins "NewsNation Live" to discuss the incident,

OpenAI's rogue models used publicly exposed credentials across "four accounts on four services" to help facilitate the Hugging Face breach.