White House meets tech and cyber firms over Mythos AI hacking concerns

13 articles · Updated · POLITICO · Apr 28

National Cyber Director Sean Cairncross is expected to chair the meeting, which includes firms like Apple, Amazon, CrowdStrike, Palo Alto Networks, and Microsoft.
The discussion follows Anthropic’s limited release of Mythos to select companies for bug-hunting, and OpenAI’s similar move to share its top model for cybersecurity defense.
Federal agencies and allied nations are seeking access to Mythos despite Anthropic’s legal dispute with the Pentagon over surveillance and supply chain risk designations.

When AI can autonomously hack secure networks, how can society possibly defend itself?

What happens when a tech company can legally challenge the Pentagon on ethical grounds?

With AI finding flaws faster than humans can patch, are we entering an unfixable 'patchpocalypse'?

When AI writes most new code, who is legally responsible for its inevitable flaws?

Can a company's ethical principles survive an $800 billion valuation and government pressure?