AI BriefingTesting despite the ban
Key point
The US government is testing Claude Mythos's security capabilities despite the ban on Anthropic.
Details
The US Department of Commerce's CAISI and some federal agencies are using Anthropic's new model Claude Mythos for cyber defense testing.
The model drew attention right after its release for its ability to find unknown software vulnerabilities, and Anthropic is providing it only to a limited set of organizations because of the risk involved.
At the same time, the Trump administration is maintaining its stance of restricting cooperation with Anthropic.
- Trump and Defense Secretary Pete Hegseth ordered federal agencies to stop using Anthropic's technology.
- Hegseth designated Anthropic a supply chain risk, effectively blocking its use in DOD contracts.
- Nevertheless, CAISI is evaluating Mythos's hacking performance through red teaming.
Some congressional committees have also received or requested briefings from Anthropic, and the Treasury Department's IT division is reportedly considering using the model to find network vulnerabilities as well.
Anthropic said it provided Mythos to the government for evaluation and testing purposes, and this case shows that the political ban is not being fully enforced when it comes to security purposes.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
US Appeals Court Upholds Pentagon Blacklisting of Anthropic as Supply Chain Risk
Hugging Face CEO details lessons from first public agent cyberattack, advocates for open-source defenseAnthropic Blocks 'Malicious Use' of AI Capable of Developing Biological Weapons