Khanna to introduce AI safety bill with ban on 'recursive' technology until safeguards exist
Rep. Ro Khanna is introducing a bill to regulate artificial intelligence, including a ban on self-improving systems until safety measures are established.
Rep. Ro Khanna is set to introduce legislation aimed at regulating the development and deployment of artificial intelligence (AI). The bill, titled the "Human Control Over AI Act," seeks to ensure human oversight and safety amidst the rapid advancement of AI technology.
A key component of the proposed legislation is a ban on what are termed 'recursive' AI models. These are systems capable of continuously self-improving or autonomously modifying their core objectives, safety mechanisms, or shutdown controls. The ban would remain in effect until federal guardrails are established and approved.
The bill emerges amid growing concerns that AI technology could evolve beyond human control. Top executives in the field have warned of potential risks, prompting Congress to establish legislative frameworks for oversight.
The proposal also includes the creation of a new federal agency to oversee advanced AI models and establish a licensing system for model training and deployment. It would also mandate extensive liability insurance for companies before releasing models and aims to deter the development of dangerous AI systems by foreign adversaries.