AI Executive Orders from 2019 to 2025
What principles, which AI ethics, have shaped U.S. federal policy on artificial intelligence over the last three presidential administrations?
This legal research project applies an ethics-mapping methodology to code the principles articulated in 12 presidential executive orders on AI issued between 2019 and 2025. It reveals a relative consensus on seven bipartisan principles of accountability, accuracy, beneficence, justice, privacy, safety and security, and trust.
These principles were first introduced by President Trump in executive orders issued during his first term, from 2019 to 2020. President Biden then reinforced these principles in his AI executive orders from 2023 to 2025. Upon his return to the White House, President Trump revoked President Biden’s executive orders on AI, and the eight AI executive orders he issued in 2025 reveal that his previous ethical commitments changed substantially.
In examining the dramatic shifts in executive power and the changing political landscape shaped by the rise of generative AI, the data reveal four significant legal findings.
- First, seven principles were affirmed in the initial executive orders on AI issued by Presidents Trump and Biden, forging a bipartisan pathway for understanding AI regulation.
- Second, these principles were either absent, substantially redefined, or explicitly rejected in President Trump’s 2025 executive orders.
- Third, President Trump used these 2025 orders to introduce novel ethical positions that fall outside traditional AI ethics frameworks.
- Fourth, the contradictory and controversial legal and ethical positions taken by the same president across two nonconsecutive terms present a unique opportunity to study the dynamic and fragile nature of AI ethics and law.
AI Executive Compact
In seeking to embody the practice of solutions-based scholarship, this research project introduces the “AI Executive Compact” (ACE), a proposed bipartisan legal framework that identifies areas of ethical consensus. The proposed compact provides evidence of common ground and illustrates how to restore enduring ethical and constitutional commitments.
The primary outcomes of this project include a comparative dataset mapping AI principles across U.S. executive orders, a policy report, and a bipartisan framework to inform and structure policy debates about both federal and state legislation.