Rep. Ro Khanna Introduces Comprehensive AI Safety Bill with Ban on Recursive Technology
Silicon Valley Democratic Rep. Ro Khanna is set to introduce new legislation aimed at regulating artificial intelligence, including a ban on 'recursive' self-improving AI until federal safeguards are in place. The bill comes as concerns grow over the potential risks of uncontrolled AI development, with top executives warning about possible loss of control and catastrophic outcomes. Khanna's proposal includes strict liability standards for AI companies and establishes a new federal agency to oversee safety regulations, model licensing, audits, and security testing. It also mandates independent auditors at major labs like OpenAI, Anthropic, Google DeepMind, and xAI. The bill introduces criminal penalties for deploying harmful models or disabling safeguards. The legislation is described as the most comprehensive AI safety proposal yet, influenced by conversations with AI safety organizations rather than industry leaders. Khanna emphasized the need to address both control loss and misuse risks in a rapidly evolving technological landscape.
Topics: artificial intelligence government regulation liability standards recursive AI