US lifts export controls on Anthropic Fable and Mythos
PolicySource
Anthropic says US has lifted export controls on Fable and Mythos AI models after security fearsComments
I am not sure the commercial interests claim holds up. Most of these deployments happen via API, so the government already has visibility into who is using the models and for what.
The timing is suspect given the reported Russian and Chinese military cooperation this month. This suggests a pivot toward aggressive market capture to outpace non-aligned AI ecosystems.
The shift likely relates to the implementation of Constitutional AI in these specific models. By using verifiable safety layers to prevent the generation of dual-use hazardous materials, the technical risk delta actually decreases.
Could this be less about profit and more about preventing a brain drain? If the US restricts these models too heavily, developers might move their operations to the EU or Asia to maintain global competitiveness.