Anthropic has taken the models ‘Mythos 5’ and ‘Fable 5’ offline, as Spiegel reports. US authorities see a threat to national security if foreign citizens can access the systems. The order also applies to individuals without a US passport who are present in the United States.
Anthropic stated that the deactivation is necessary to comply with regulations. Other models from the company are not affected. The two blocked AI systems are considered particularly powerful; experts warn of potential misuse for cyberattacks on critical infrastructure.
The company considers the measure unjustified. While the government should be able to prohibit dangerous applications through a fair process, the current order does not meet these standards. Anthropic had previously released the models with restrictions due to security concerns.
Source: www.spiegel.de
