Khanna bill would pause recursive AI

Rep. Ro Khanna will introduce the Human Control Over AI Act, a proposal that would ban models from recursively improving themselves or changing core objectives, containment, or shutdown controls until federal safeguards and agency approval exist. The Next Web’s summary says the bill would create one federal agency to oversee models from 4 named labs—OpenAI, Anthropic, Google DeepMind, and xAI—with licensing, audits, and safety standards, while no House vote is expected before the midterm election. For builders, file this as an early liability-and-release-bar signal rather than current law: the archive’s AISI evaluation analysis shows why scope controls and independent evaluation are already becoming deployment requirements.