Congressman Introduces Bill to Regulate AI Development

Safeguarding Against Unchecked AI

Representative Ro Khanna is proposing new legislation aimed at controlling artificial intelligence. The bill, called the Human Control Over AI Act, seeks to establish clear rules for AI development. It was introduced on September 28, 2026.

The act focuses on two main areas. It would impose strict liability on AI developers. It also proposes a ban on AI systems that can improve themselves recursively. This ban would remain until government safety measures are in place.

Khanna's bill addresses growing concerns about the rapid advancement of AI technology. The legislation aims to ensure that humans maintain oversight. This is especially important for AI systems that could evolve beyond our understanding or control.

When Can AI Improve Itself?

The proposed strict liability means companies developing AI would be held accountable for any harm caused by their systems. This could include financial damages or other penalties. Such a measure is designed to incentivize developers to prioritize safety and ethical considerations.

The ban on recursive self-improvement is a key component. This refers to AI that can modify and enhance its own code without direct human intervention. Proponents argue that such systems could quickly become unpredictable. The bill seeks to prevent this until robust regulatory frameworks are established.

Frequently Asked Questions

This legislation comes at a time when AI capabilities are expanding dramatically. Discussions about AI's potential risks and benefits are ongoing globally. Khanna's bill represents a significant legislative effort to proactively manage these risks.

What is strict liability in this context? Strict liability means AI developers can be held responsible for damages caused by their AI, regardless of fault. This encourages greater caution in development.

What is recursive self-improving AI? This is AI that can independently rewrite and upgrade its own programming. The bill seeks to prevent its use until safety standards are met.