Anthropic, a US-based AI startup, said in a July 27 post by CEO Dario Amodei that it does not support a ban on open-weight AI models as a category, including Chinese ones, as Washington scrutinizes their use by US companies.

US lawmakers are investigating Chinese-developed AI models used by US companies.

Policy experts said possible restrictions could include federal procurement bans affecting government agencies and contractors.

Anthropic instead called for keeping powerful chips and chipmaking equipment out of China, curbing industrial-scale model distillation, and requiring safety testing for sufficiently capable models, whether open or closed.

Open-weight models generally mean releasing weights so others can download, run, and fine-tune the models.

They usually do not include the code, checkpoints, or data needed to fully reproduce development.

US semiconductor controls on China began in 2022 and were later tightened.

Some analysts are now pushing stricter limits after a January 2026 rule permitted sales of some advanced AI chips to China.

Distillation is also emerging as a policy issue.

The White House warned in April about industrial-scale campaigns by Chinese entities to distill US frontier models, and a House bill aimed at deterring unauthorized distillation has advanced.

🔗 Source: Anthropic

Recent Anthropic developments