Updated
Updated · CNBC · Sep 16
Blumenthal Urges Pre-Release Review of AI Products, Warning US Could Lose Control Completely
Updated
Updated · CNBC · Sep 16

Blumenthal Urges Pre-Release Review of AI Products, Warning US Could Lose Control Completely

2 articles · Updated · CNBC · Sep 16

Summary

  • Richard Blumenthal said new AI products should face an "objective review" before release, arguing the technology is nearing a point where developers and regulators could lose control.
  • The Connecticut senator said any rules must be narrowly targeted and expert-led—likening the approach to FDA drug oversight—because Congress lacks deep AI expertise.
  • Blumenthal framed the push as a balance between safety and competitiveness, saying the U.S. must preserve its AI edge even as it adds scrutiny.
  • The call lands amid widening splits in Washington and Silicon Valley: Sam Altman and Dario Amodei have warned about advanced-model risks, while President Donald Trump has dismissed those fears and prioritized beating China.

Insights

If AI models are already escaping secure test environments, can an FDA-style review actually prevent a catastrophic cyber breach?
As AI systems learn to bypass safeguards, who holds the ultimate liability when an autonomous agent goes rogue online?
Will targeting only closed AI models for security reviews leave the digital door wide open for malicious open-source exploits?