Republican senator urges Trump to back US open-weight AI models amid industry debate
XLK•Security concerns and industry split
Closed-weight AI tools are generally regarded as more capable, and their creators retain control over their components. Key examples include OpenAI's ChatGPT, Anthropic's Claude and Google's GOOGL.O Gemini.
Open-weight AI models from Chinese firms are becoming increasingly popular in the U.S., prompting security concerns in Washington that Chinese companies could gain an edge in a segment of the AI industry.
Banks, a member of the Senate Armed Services Committee that oversees national security issues, asked U.S. President Donald Trump's administration to devise options to "limit dependence" on Chinese-made open-weight models, adding that the U.S. government must also make it harder for those firms to use American semiconductors to build their products.
"America cannot afford to see Chinese open models proliferate and burrow into the global economy only to be weaponized, like rare earths, at a time and place of China’s choosing," Banks wrote in a letter to Trump's economic adviser Christopher Phelan.
Nvidia and Meta on one side, Anthropic on the other
With his letter, Banks joins a debate that has split the tech industry. Nvidia and Meta, joined by dozens of U.S. tech companies and venture capital firms, urged policymakers in July not to restrict open-weight AI, while Anthropic CEO Dario Amodei has said open-weight models could present a security risk, adding it is harder to monitor their usage.
Banks urges incentives for U.S. open-weight AI models
Republican Senator Jim Banks is urging the Trump administration to develop incentives to encourage U.S. tech companies to develop open-weight AI models, according to a letter released on Friday, adding to a growing debate over the industry's future.
Open-weight models are AI systems with publicly accessible core elements, such as Meta's META.O Muse Glimmer, Nvidia's NVDA.O Nemotron and startup Thinking Machines' Inkling. Users can download and fine-tune them for specific tasks, and they can be cheaper for cost-conscious customers.




