The White House instructed U.S. government agencies considering new rules to “conduct risk assessment and cost-benefit analyses” and ensure “flexible frameworks rather than one-size-fits-all regulation.” They are required to ensure AI is fair, non-discriminatory, transparent, safe and secure. (Reuters)

Talking point: The administration’s principles are designed to carve out a distinct U.S. position on AI development and use as the Trump administration tries to establish and maintain American control over the sector. The U.S. has refused to join its G7 peers in the Global Partnership on AI (GPAI), a Canada- and France-founded forum to ensure the technology’s growth takes “a human-centric and ethical approach” and that users consider diversity and sustainability among other high-level principles. The group’s decisions aren’t binding, but U.S. officials still think it could be too restrictive, and want countries to set their own rules. Ottawa is doing both—participating in the GPAI and establishing its own rules for automated decision-making within government.