AI Model Cracks Weakened Encryption Standards

An AI model called Claude Mythos Preview, created by a company called Anthropic, has found weaknesses in a less powerful version of a widely used internet security system called AES.

This system is important for protecting things like bank transactions and secret information.

The AI also successfully attacked another security system called HAWK, which is supposed to be resistant to attacks.

The findings were shared with the US government and industry partners before being made public.

While the current internet security systems are still safe, this research shows that AI could pose new challenges to digital security in the future.

The US government is also paying more attention to AI after some incidents raised concerns.

There is also a competition between the US and China to develop advanced computers that could break existing encryption standards.