A proposed "People's AI Constitution Council" would consist of a randomly selected body of 100 citizens from across the country, serving two-year terms, to address the lack of public oversight over AI models adopted by the government. Prior to their term, Council members would undergo a three-month briefing on AI fundamentals and constitutional AI, with training materials compiled by AI experts in government and industry, likely involving the Center for AI Standards and Innovation (CAISI). The Council would select—not draft—three constitutions from a pool drawn from frontier AI labs, research organizations, and public submissions with at least 100,000 electronically gathered signatures. CAISI would provide objective analysis of submissions to ensure they lack key components or pose undue risks, but would not select constitutions. Access to federal contracts would be conditioned on a model adopting one of the Council’s constitutions, making it a procurement standard. Selected constitutions would remain in place for two years, subject to CAISI review, which could convene the Council to suggest sample amendments if frontier AI capabilities render a constitution inadequate. Subsequent Councils could retain, amend, or replace prior constitutions.