Democratic Congressman Ro Khanna is planning a bill that would ban recursively self-improving AI systems until federal safety regulations are in place. CNBC reports this.
The proposal, called “Human Control Over AI Act”, would also create a federal AI agency responsible for licensing, audits and safety requirements for advanced models. Independent reviewers would be placed at AI laboratories, while companies would be subject to strict liability rules and requirements for extensive liability insurance.
The bill also includes penalties for people who knowingly circumvent safety or shutdown systems.
