Rep. Ro Khanna is set to introduce a significant bill aimed at regulating artificial intelligence, reflecting increasing anxiety among lawmakers and industry leaders about the potential dangers of AI technology. The 'Human Control Over AI Act' proposes strict liability standards and a prohibition on recursive self-improving AI until the government can implement necessary safety measures.
Khanna emphasized the existential risks associated with AI, stating that there is a 'civilizational extinction risk' if the technology is not properly controlled. The bill would establish a new federal agency responsible for overseeing AI models from major companies like OpenAI, Anthropic, Google DeepMind, and xAI, focusing on safety regulations, licensing, and continuous testing.
Additionally, it would require independent auditors at frontier labs and impose legal penalties for AI-related failures, including criminal charges for actions that compromise safety systems. This legislation is part of a broader movement to ensure that AI development is guided by safety considerations rather than solely by corporate interests.
However, no votes on this or other AI safety proposals are anticipated until after the upcoming midterm elections