Anthropic's Mythos AI sparks alarm

The Bank of England has announced it will include AI-driven risks in its upcoming financial system stress tests, specifically citing the threat posed by advanced models like Mythos.

While Anthropic has limited the model's release to prevent hackers from sabotaging networks, the U.S. government is establishing safeguards to allow the Department of Defense and Treasury to leverage its capabilities.

Security researchers have already identified a potential 'design flaw' in the software that could put 200,000 servers at risk of takeover.

The debate intensifies as tech giants lobby for secrecy regarding the environmental and security tolls of these massive data centers.