What should the Federal government do today to ensure America succeeds with advanced AI in the future? What foundation should we lay in the short term—not with the tools we might dream up, but with the tools we have in front of us now—to bring about flourishing in the long term?
Today, the Trump Administration releases the AI Action Plan, and these are the questions we set about to answer. It is authored by Office of Science and Technology Policy Director Michael Kratsios, AI and Crypto Czar David Sacks, and Secretary Marco Rubio, in his capacity as Assistant to the President for National Security Affairs. This plan is the culmination of months of research, writing, and collaboration that took place across the Federal government. It is my pleasure to share a bit about it with you here.
The plan catalogues a wide variety of discrete and actionable steps the Federal government can take with existing funding and statutory authorities across its many agencies to advance American AI in the coming 1-3 years. It does so across three broad pillars: innovation, infrastructure, and diplomacy and security. In the innovation pillar, we focus on removing red tape that harms AI development or deployment, as well as fostering transformative adoption of AI throughout the economy and government. In the infrastructure pillar, we present a roadmap for the construction of the vast AI infrastructure that will be needed to power AI. And in the final pillar, we articulate a strategy for international engagement to both promote American AI with allies and partners, and to prevent adversaries from misusing the technology to harm Americans.
We look forward to operationalizing this plan over the weeks, months, and years to come, a process that will begin in the coming days.
On a personal note, I’d like to thank the many people across the Federal government whose effort made this plan possible. The interagency process—the process by which relevant parties within the Federal government provide feedback on documents such as the Action Plan—is often maligned in DC. This criticism, I think, is misplaced. The input we received from agencies consistently sharpened, refined, and strengthened the Action Plan. Without it, the document would be far weaker. Throughout the process, the depth and breadth of talent and knowledge I encountered among my colleagues has astounded me.
I’d also like to thank the thousands of people—groups and individuals alike—who submitted public interest comments in our Request for Information on the Action Plan. This public input, too, was essential.
With this Action Plan, we have endeavored to meaningfully advance the ball on many of the core issues of AI policy—and many others that perhaps get less attention than they should. We have tried to think about what can be done in the near term to advance American interests in the long term. And though it may be obscured somewhat by the acronyms and the technical terminology, I assure you we have taken to our task with zeal, care, and love for our country.
I hope you enjoy reading it.
-Dean