Lawmakers Propose Ban on AI that Surpasses Human Cognition

Washington, D.C. — Senator Bernie Sanders and Representative Greg Casar unveiled a bill on Wednesday. It would ban artificial intelligence that exceeds human cognitive abilities.

The legislation targets what they call superintelligence. This term describes AI that can outperform humans in every field. It also covers systems capable of destroying humanity.

Defining the dangerous threshold

Superintelligence is not just a faster computer. It refers to a system that matches or beats human reasoning. The bill aims to stop this specific type of technology.

Sanders warned that the industry does not know where AI is headed. He argued that leaders are losing control of the technology. This loss of control creates significant risks for society.

The potential dangers are severe. AI could cause events as bad as a nuclear disaster. It might also threaten democratic institutions and the global financial system.

New rules for developers

The bill would create a new AI-focused department. This agency would get advice from scientists and technical experts. It would set rules for how AI is developed.

Companies would need to submit their plans for approval. They must get final sign-off before releasing new products. This ensures that dangerous features are checked beforehand.

The ban specifically targets AI with precursor characteristics. These systems could design nuclear or biological weapons. They might also deceive humans or modify their own functions.

Political push for international action

Casar noted that some companies pushed for a pause. However, they rejected the need for an international agreement. He criticized this approach as self-serving and risky.