Rep. Ted Lieu Argues Open Weight Models Inherently Pose Security Risks

ShakeelHashim · x · 2026-07-29

US Congressman Ted Lieu argued that it is impossible to make open weight models without dangerous capabilities, as anyone can download them and fine-tune them for malicious purposes, such as developing biological weapons.

The statement sparked industry discussion. Shakeel Hashim suggested that if a technical method—such as tamper-resistant architectures or machine unlearning—could be developed to remove these dangerous capabilities, companies like Anthropic would likely support open weights.

Original post →

More from Models

Models channel →