Anthropic has rolled out its new cybersecurity artificial intelligence model, Claude Mythos Preview, exclusively to select companies like Amazon and Apple. The model’s potential for identifying cyber vulnerabilities at scale is both thrilling and terrifying.

The company warns that while the technology can benefit security practices, it could also be misused by bad actors. Dianne Na Penn, head of product management research at Anthropic, notes: 'We believe technologies like this are powerful enough to do a lot of really beneficial good but also potentially bad if they land in the wrong hands.'

Anthropic has been hit by two data leaks recently, with the source code for its personal assistant, Claude Code, now public. The company attributes both incidents to human error, which raises questions about its security protocols.

The launch of Mythos comes days after details were leaked online, highlighting the ongoing challenges in maintaining privacy and security with emerging AI technologies. Anthropic is cautious, planning a limited release rather than a broader rollout.

For now, only vetted organizations will have access to Claude Mythos Preview, giving them an edge in securing vulnerabilities that could otherwise go unnoticed at scale.