Anthropic Blocks Access to New AI Models Following US Government Order

US artificial intelligence firm Anthropic has suspended access to its advanced AI models, Fable 5 and Mythos 5, for all users globally. This move comes in response to a directive from the US government, which mandated that the models should not be accessible to non-US citizens, citing national security interests.

The government's order specifically targeted individuals without US citizenship, including those residing in the US and foreign employees of Anthropic. In reaction, the company decided to implement a complete block, affecting all users. The AI behind Mythos 5 is reportedly adept at identifying software vulnerabilities, including those that have remained undiscovered for decades. While this capability can help organizations patch security flaws, there are significant concerns about its potential misuse for cyberattacks.

“

Anthropic stated that it had only received incomplete information from the government so far.

Mythos 5 is the unrestricted, non-public version intended for use by government agencies and select companies to enhance their cybersecurity. Fable 5, released recently, is based on the same technology but incorporates enhanced safety features, particularly in areas like cybersecurity and biotechnology. Anthropic has expressed that the government's order is disproportionate, stating they received incomplete information and that the identified risks are not unique to their models. The company asserts that Fable 5's safety protocols have undergone thorough testing.

“

The company believes the block is excessive, as similar risks exist with models from other providers.