“Information about current AI capabilities, future projections, and potential risks is heavily concentrated within the AI companies themselves, creating a significant information asymmetry with the public and regulators.”