President Trump's Plan to Fully AI America

“The United States is in a race for global dominance in artificial intelligence,” reads the introduction to the White House’s AI Action Plan released on July 23. The main competitor in this race, as other passages have made clear, is China.

According to US President Donald Trump, the action plan clearly frames AI as a competition between great powers, where whoever wins takes all. Phrases like "ALL-IN" or "Winning The AI race" were decorated throughout the announcement ceremony, showing that for the US, this is an "all-in" race and the winner is required.

President Donald Trump signs into law the $500 billion Comprehensive AI Plan for America. Photo: Washington Post

“Whoever has the largest AI ecosystem will set global AI standards and reap enormous economic and military benefits,” the introduction continues. “Just as we won the space race, it is imperative that the United States and its allies win this race.”

While most of the specific items in the Action Plan relate to domestic research and development (R&D), the economy , or international trade, it also includes a long list of to-dos for the Pentagon.

Of the roughly 19 “recommended policy actions” cited explicitly from the Department of Defense , five relate to the internal workings of the world’s most powerful military.

The remaining items all give the Pentagon a prominent, even leading, role in setting standards and coordinating efforts across the executive branch.

Particularly interesting is the call for DARPA to lead an interdisciplinary “technology development program” to address a fundamental problem with advanced AI, ignorance.

This is especially true for generative AI like Large Language Models, which can summarize large volumes of data and even generate surprising insights but also frequently “hallucinate” and misinform in unpredictable ways.

“Technologists understand how LLM works, but often cannot explain why a model produces a particular result,” the Action Plan notes.

This lack of predictability… could make it difficult to use advanced AI in defense, national security, or other applications involving human life. This makes GenAI a difficult tool to trust even at lower risk, as a recent Marine Corps field manual warned.

A pilot with the 149th Intelligence Squadron conducts training at the computer training laboratory at Mather Field, California. Photo: US.DOD

In addition to the DARPA-led program, the task force — titled “Interpretability, Control, and Robustness of AI” — also calls on the Pentagon to collaborate with the departments of Energy, Commerce, Homeland Security, and others on “AI hackathons.”

Essentially an open invitation to outside experts to conduct (non-destructive) attacks on their systems and report what they find. The Pentagon began adopting this approach with its “Hack the Pentagon” competition in the final year of the Barack Obama administration.

Another Pentagon initiative promoted in Trump’s plan today adopts the slogan “responsible AI.” One of Trump’s former defense secretaries, Mark Esper, was the one who formally directed the Defense Department to follow a list of ethics principles, highly specific interactive online checklists for officials running AI programs.

The AI Action Plan recommends that federal agencies, “led by the Pentagon,” should “continue to refine the Pentagon’s Responsible AI and Generative AI Frameworks, Roadmaps, and Toolkits.”

The Action Plan also gives the Pentagon a leading or prominent role in a number of AI security initiatives. The Pentagon should lead interagency efforts to “develop new technical standards for high-security AI centers” and assess adversaries’ progress in AI.

The Defense Advanced Research Projects Agency (DARPA) is assigned many key roles in President Trump's AI Plan. Photo: US.DOD

The Department of Defense will support the Office of the Director of National Intelligence in developing “AI Assurance Standards” for the Intelligence Community, which is largely staffed at the Pentagon. The agency will also work with other agencies on issues ranging from promoting domestic high-tech manufacturing to improving export controls.

Of the five recommendations for upgrading the Pentagon, perhaps the most intriguing — albeit vague — is the call to “establish a Virtual Testbed for AI and Autonomous Systems.”

It's not even clear whether this is a real-life testing facility or some kind of Matrix-style super-simulation: The "virtual" in the name implies it's all online, while the call to evaluate "the geographic requirements… necessary for such a facility" suggests there could be a physical location.

The Pentagon has also been tasked with turning military universities into “AI research hubs,” upskilling the workforce and defining “workflows.”

Most worryingly, the Action Plan also wants the Pentagon to have priority access to cloud providers and other computing services so it can continue to use them during wartime.

Original article link Copy link

https://breakingdefense.com/2025/07/how-the-pentagon-plays-into-trumps-sprawling-artificial-intelligence-action-plan/

Source: https://khoahocdoisong.vn/ke-hoach-ai-hoa-toan-dien-nuoc-my-cua-tong-thong-trump-post2149040728.html