US administration and Anthropic in talks over Mythos AI

At the center of the discussions closely watched by the tech world are recent contacts between Anthropic CEO Dario Amodei and White House Chief of Staff Susie Wiles. These talks concern whether the new artificial intelligence model known as Mythos will be made accessible to the American government. Criticism from the government and lawsuits filed in federal courts against Anthropic have increased, particularly due to the risks the model may pose to national security.

The significance of Mythos stands out not only for its technological innovations but also for its potential security vulnerabilities. Officials from the US Department of the Treasury and other official agencies state that the model is notable for its ability to detect and exploit cybersecurity vulnerabilities. For this reason, some officials are pressuring the White House to have federal agencies conduct comprehensive testing of Mythos.

As a result of the discussions, the White House administration has not yet reached a definitive decision on whether to give the green light for federal agencies to actively use Mythos. In contrast, many high-level figures, led by US Secretary of Defense Pete Hegseth, emphasize that the security risks must be balanced against the benefits to be gained from artificial intelligence.

The Anthropic camp is taking a more cautious approach. In a statement released by the company, it was noted that the Mythos model would not be opened for general access until security tests are completed and it is fully cleared of potential risks of misuse.

During this process, recent power outages in Anthropic's technical infrastructure and limited processing capacity are also cited as primary issues that must be resolved before the model's large-scale release. Sources close to the company state that Mythos will not be offered for widespread use until there is certainty that it will provide stable service.

News Source: 12punto