US government halts Anthropic AI models over security concerns

Image credit: Search Engine Journal

The United States government on Wednesday issued an export control directive to Anthropic, compelling the company to suspend access to its Fable 5 and Mythos 5 artificial intelligence models citing national security concerns.

The directive, issued at 5:21 p.m. EST, mandates that Anthropic block all foreign nationals, including employees, from accessing the specified AI models, leading the company to disable access for all customers globally.

Government officials believe a method exists to bypass safety guardrails within the Fable 5 model, a vulnerability Anthropic reportedly considers minor.

Anthropic, however, contradicted the government’s assessment of its cybersecurity posture. The company stated it has strong safeguards in place and employs a defense-in-depth strategy for its Fable 5 model.

The move highlights an ongoing dispute between Anthropic and the U.S. government. Anthropic has previously resisted allowing its products for mass domestic surveillance and fully autonomous weapon systems, according to sources familiar with the matter.

The export control directive specifically targets access for any foreign national, regardless of their employment status with Anthropic. This broad scope prompted the company to implement a complete shutdown of access to both Fable 5 and Mythos 5 for all users to ensure compliance.

The government’s order underscores escalating tensions between regulators and leading AI developers regarding the control and potential misuse of advanced artificial intelligence technologies.

Anthropic has not yet indicated when or if access to Fable 5 and Mythos 5 might be restored, pending resolution of the national security concerns raised by the U.S. government.

Source: Search Engine Journal