トランプ氏のAI命令:自主的な監督か、それとも規制の抜け穴か?
トランプ大統領が、フロンティアモデル向けに自主的な30日間のリリース前審査を設ける新たなAI大統領令に署名。この政策は監督体制への転換を示すものの、必須ライセンス制度が欠けており、企業が安全チェックを回避する余地を残している。
less than two weeks after rescinding a previous ai executive order, president trump introduced a new policy requiring tech companies to voluntarily submit frontier models for government review 30 days before release. the order establishes a dedicated ai cybersecurity clearinghouse to coordinate security checks and explicitly rejects mandatory licensing for model deployment. this 30-day window is significantly shorter than the 90-day requirement in trump's prior order, indicating a compromise between innovation interests and security concerns. critics argue the voluntary nature of the review creates a loophole that major labs could exploit, undermining the policy's effectiveness. the move marks a strategic pivot from a purely hands-off approach, setting the stage for a brewing battle over ai regulation in the us.