On December 11, 2025, President Trump signed his long-anticipated Executive Order (EO) on “Ensuring a National Policy Framework for Artificial Intelligence,” which claims to offer a federal policy framework to preempt state-level AI laws across the country. 

The most concerning aspect is the creation of an AI Litigation Taskforce, which would be in charge of suing states that have enacted AI laws that are deemed “inconsistent” with the vaguely worded policy set out in the EO. The taskforce will act according to the discretion of Attorney General Pam Bondi, in consultation with David Sacks.

Second, the EO directs the Special Advisor for AI and Crypto (David Sacks) and the Assistant to the President for Science and Technology (Michael Kratsios) to prepare a legislative framework to preempt state laws. Notably, the EO carves out certain areas that would not be subject to preemption, including state laws related to child safety, AI compute and data center infrastructure, and state government procurement and use of AI.

The Trump Administration will now attempt to enforce the directives defined in the EO in the initial 30- and 90-day windows outlining the formation of the AI Litigation Task Force, evaluation of preexisting state laws, issuance of a policy directive potentially threatening BEAD funding for certain states found to have laws inconsistent with the EO, determination around adopting a federal reporting and disclosure standard, and assessing potential applicability of the Federal Trade Commission Act’s deceptive conduct authority to override state laws related to algorithmic bias. 

In practice, the Executive Branch cannot overturn state laws that have already been passed by state legislatures and signed into law by governors. The AI Litigation Taskforce seems to challenge state laws under the “dormant commerce clause” theory, which is rarely successful. We expect widespread legal challenges to the EO by state attorney generals, with additional advocacy from outside organizations only adding to the pressure. Additionally, at the federal level, once the Trump Administration engages in any actual implementation steps – whether that be initiating a federal rulemaking process or otherwise – there are a number of options at one’s disposal, such as triggering a CRA, or, in other words, a method for Congress to utilize the Congressional Review Act to overturn a specific review or directive of a federal government agency. Further, at the state level, the efforts of state legislators working to enact legislation meaningfully addressing the short and long-term harms posed by AI systems will be put at risk.

Since the signing of the EO, a plethora of federal and state legislators, journalists, and organizations have written responses in opposition to the EO, a selection of which are included below:

• Don Beyer, U.S. Representative, VA-08, denounced the EO in a X post and press release.
• Doris Matsui, House E&C Communications and Technology Subcommittee Ranking Member, denounced the EO in a X post and press release.
• Zoe Lofgren, House SST Committee Ranking Member, denounced the EO in a press release.
• Sara Jacobs, U.S. Representative, CA-51, denounced the EO in a X post and press release.
• Adam Schiff, U.S. Senator from California, wrote in opposition to the EO in a X post.
• Alex Bores, NY State Assemblymember and candidate for U.S. Congress in NY-12, denounced the EO in a X post.
• Steve Bannon, ex-Trump adviser, sent a text message to Axios stating that David Sacks had “completely misled the President on preemption.”
• Will Steaken, Senior Washington Reporter, ABC News, outlined the growing backlash from part of the MAGA base on stripping state’s rights on regulating AI, in a X post.
• ACLU statement denounced the EO via a press release.
• Office of California Governor Gavin Newsom denounced the EO in a press release.
• Huo Jingnan, a Reporter at NPR, wrote about the potential illegality of the EO in an article.

The final EO signed by the President was largely similar to the EO draft text that was released a couple of weeks prior. A section-by-section comparison of those two different EO texts are included below.

Section by Section EO Comparison

Section 1. Purpose

With the exception of some specific language changes, the most notable change in the final version of the EO was the exclusion of a reference to California SB 53, which was described in the original EO as a “complex and burdensome disclosure and reporting law premised on the purely speculative suspicion that AI might ‘pose significant catastrophic risk.’”

In one other change, the signed EO took out language from the original version that identified “sophisticated proponents of a fear-based regulatory capture strategy [that] are responsible for inciting [different patchwork] laws.”

Section 2. Policy

Both versions remained effectively the same.

Section 3. AI Litigation Task Force 

Both versions remained effectively the same.

Section 4. Evaluation of State AI Laws 

Both versions remained effectively the same.

Section 5. Restrictions on State Funding 

Both versions remained largely the same, with the exception of adding the Special Advisor for AI and Crypto (David Sacks) to serve in a consultation role with executive departments and agencies to assess discretionary grant programs under Subsection (b). 

Section 6. Federal Reporting and Disclosure Standard 

Both versions remained effectively the same.

Section 7. Preemption of State Laws Mandating Deceptive Conduct in AI Models 

Both versions remained effectively the same.

Section 8. Legislation 

In Subsection (a), the role of preparing a AI federal policy framework proposal was altered from the Special Advisor for AI and Crypto (David Sacks) and Director of the White House Office of Legislative Affairs (James Braid) to David Sacks and the Assistant to the President for Science and Technology (Michael Kratsios).

The entirety of Subsection (b) was added in, outlining potential carveouts for state law preemption related to child safety, AI compute and data center infrastructure, and state government procurement and use of AI.

Section 9. General Provisions 

Both versions remained effectively the same.