Open-source AI models could face regulatory bans within six months

U.S. regulators are debating new executive orders that could restrict open-source artificial intelligence models, with implementation potentially occurring within six months. These discussions respond to the advancement of Chinese open-source models, particularly DeepSeek, which demonstrate capabilities matching the most sophisticated systems currently available. Anthropic has taken a prominent role in promoting regulatory measures, citing security concerns and potential extraction of proprietary technologies. The company released reports on unauthorized foreign access to its systems and submitted policy recommendations to lawmakers. Critics contend this activity constitutes regulatory capture, where Anthropic gains economic advantage while restrictions would harm the emerging U.S. ecosystem of companies dependent on open models. The analysis argues that unilateral bans would be ineffective without global coordination. It calls for Microsoft and Meta to release open-source models with comparable capabilities to shift the focus of the discussion. The absence of coherent security solutions for commercial APIs undermines arguments specifically targeting open-source models, and meaningful resolution requires international collaboration rather than restrictions imposed solely by the United States.