Arcee CTO says Chinese open-weight models pose no security risk
A U.S.-based AI lab, Arcee, says there's no need to worry about Chinese open-weight AI models being a security risk.
According to its chief technology officer, Lucas Atkins, using these models in business settings keeps them just as secure as any other open-source tech.
This comes as people debate the safety of affordable Chinese AI like Alibaba's Qwen and Moonshot's Kimi K3.
Lucas Atkins backs open-source model releases
Atkins points out that while how these models are trained stays private, organizations can still check and tweak the source code for their needs.
He believes the US should focus on innovation and teamwork instead of banning foreign AI.
With more companies choosing flexible, model-agnostic approaches, Atkins thinks releasing better open-source models is key for staying ahead in global AI.