The United States lifted restrictions it had imposed on the export of Anthropic's latest artificial intelligence models more than two weeks after they went into effect.
The company announced on Tuesday that the US Department of Commerce had lifted restrictions on the export of its artificial intelligence models "Mythos 5" and "Fibel 5," confirming that it would begin restoring access to the two models starting Wednesday.
The US government had previously ordered Anthropic to suspend the availability of its AI models to foreign nationals, including company employees, prompting it to completely block the two models in compliance with the order.
The "Mythos" model is characterized by an advanced ability to detect software vulnerabilities, including vulnerabilities that may have remained undetected for decades. It has previously been used by US authorities and a number of select companies to help detect and address these vulnerabilities.
Conversely, these capabilities have raised concerns from the outset about the possibility of this type of artificial intelligence being exploited as a dangerous cyber weapon if it falls into the wrong hands.
In a letter published by US media announcing the lifting of restrictions, US Commerce Secretary Howard Lutenick said that Anthropic had agreed, among other commitments, to identify and proactively address security risks in its models.
