Arcee Says Chinese Open Models Pose No Inherent Threat

Open-weight models can be downloaded, modified and deployed independently, lowering costs and broadening access for businesses, developers and researchers. Their rise has also sharpened concerns because safeguards can be removed and released weights cannot be recalled. As Chinese laboratories narrow the performance gap with leading U.S. systems, policymakers and technology companies are debating how to preserve competition without worsening intellectual-property, cybersecurity and national-security risks.

Anthropic Chief Executive Dario Amodei said on July 27, 2026, that the company had never advocated a blanket ban and regarded open models without dangerous capabilities as a public good. He proposed three targeted measures: restricting China’s access to advanced chips and chipmaking equipment, curbing industrial-scale model distillation, and requiring mandatory safety tests for sufficiently capable open and closed models. Amodei said the gravest threat could instead be a secret model supplied to China’s People’s Liberation Army for drones or to state security agencies for surveillance.