Anthropic still blacklisted, but Mythos is a separate issue
CNBC • • business
Key Points:
- Department of Defense CTO Emil Michael acknowledged that while Anthropic remains a supply chain risk, its AI model Mythos represents a distinct national security concern due to its advanced cyber vulnerability detection and patching capabilities.
- The DOD had previously blacklisted Anthropic's Claude models, requiring defense contractors to certify non-use, leading Anthropic to sue the Trump administration to reverse this designation.
- Despite the supply chain risk status, the DOD and the NSA continue to use Anthropic's models, including Mythos, particularly in military operations related to Iran.
- The DOD recently signed agreements with seven AI companies, including Google, OpenAI, Nvidia, and Microsoft, to deploy AI technologies across classified networks, while negotiations over guardrails with Anthropic remain ongoing.
- Anthropic's CEO met with Trump administration officials to discuss Mythos, with President Trump indicating a potential deal, though Anthropic's lawsuits against the government continue.
