Anthropic suggests tweaks to proposed U.S. AI chip export controls

Anthropic agrees with the U.S. government that implementing robust export controls on domestically made AI chips will help the U.S. compete in the AI race against China. But the company is suggesting a few tweaks to the proposed restrictions.

Anthropic released a blog post on Wednesday stating that the company “strongly supports” the U.S. Department of Commerce’s “Framework for Artificial Intelligence Diffusion” ahead of the interim rule’s implementation date on May 15.

The framework was proposed by outgoing president Joe Biden in January and is meant to bolster AI chip export controls for the purposes of national security and to ensure the U.S.’ dominance in AI. It divided the world’s countries into three tiers, with each tier having its own guidelines and restrictions.

Tier 3, the most restrictive tier, which includes countries that were already impacted by existing export controls, like Russia and China, would face additional restrictions. Tier 2 countries, like Mexico and Portugal, would come under export restrictions for the first time and would have a cap on how many chips they could buy. Tier 1 countries, like Japan and South Korea, would continue without export restrictions.

When these restrictions were proposed in January, semiconductor giant Nvidia released a statement calling them “unprecedented and misguided,” and suggesting that they would “derail” innovation worldwide.

Clearly, U.S.-based AI companies, like Anthropic, don’t agree. In its blog post, the lab expressed support in broad strokes for the restrictions.

Anthropic did, however, propose lowering the number of chips Tier 2 countries can purchase without review and instead encouraging these countries to buy more chips through government-to-government agreements to avoid smuggling and increase U.S. control.

The company also thinks the U.S. government should increase funding to ensure these export controls are properly enforced.

This statement by Anthropic is not particularly surprising. Anthropic’s CEO Dario Amodei has been one of the more vocal U.S. AI leaders in favor of export restrictions. Amodei wrote an op-ed in the Wall Street Journal in January about why the U.S. needs stronger chip export controls.

TechCrunch has reached out to Anthropic for more information.