Where Is the Force in the AI Force?

When President Trump announced that he was forming an “AI Force,” my first reaction was not to ask who would lead it. I wanted to know where the force was.

Trump said he was forming the AI Force “much like” he created Space Force and would soon appoint a new AI czar. He promised that his administration would not hinder or stifle artificial intelligence. It would “cherish it, help it, and watch over it” while looking for what he called “BAD” through the existing criminal and civil justice systems.

It was classic Trump. Find the name. Own the message. Command the news cycle. Let everyone else argue about what it means.

“AI Force” is good marketing. It sounds stronger than the Interagency Council on Artificial Intelligence Risk, Development and Competitiveness. It invokes national purpose and competition with China. It reassures the industry while telling a nervous public that Washington is watching.

But a launch message is not an operating plan.

As of this writing, there is no published executive order, charter, budget, staff or chain of authority. Trump has not named the czar. The White House has not explained whether this will be a civilian organization, military organization, interagency task force or advisory council. It has not said what the force can investigate or require.

Trump has created the brand. He has not yet shown us the institution.

Marketing Can Become Machinery

With Trump, that distinction matters. He often markets an idea before the details exist. That does not necessarily mean the idea is fake or destined to disappear.

Space Force is the obvious example, which is why his comparison deserves more than an eye roll. Many people initially treated it as a slogan or punchline. It became a real military service after Congress established it through the fiscal 2020 National Defense Authorization Act, giving it a mission, personnel, leadership, appropriations and oversight.

Trump supplied the brand and political momentum. Congress and the Pentagon supplied the institutional force.

That is the standard the AI Force must meet. Space Force did not become real because Trump named it. It became real when it gained the people, money, mission and authority to act.

A Force Needs More Than a President’s Pen

There is another lesson in the Space Force story that deserves more attention. Trump did not establish a permanent new military service solely through an executive order. Congress passed the United States Space Force Act as part of the fiscal 2020 National Defense Authorization Act, and Trump signed it into law. The statute established the service, defined its composition and functions, and placed it within the Department of the Air Force.

That distinction matters.

Executive orders have their place. A president can direct executive agencies, coordinate work across the administration and exercise powers that the Constitution or Congress has already given the executive branch. But an executive order is not a substitute for legislation. When a president attempts to exercise power Congress has not authorized, the order may not survive judicial scrutiny. The Supreme Court’s longstanding Youngstown framework makes presidential power strongest when the president acts with congressional authorization and far more vulnerable when that support is absent.

An executive order is also temporary by nature. What one president creates with a pen, the next president can often revoke with another pen. Trump demonstrated that himself when he returned to office and rescinded a long list of executive actions issued by the Biden administration.

If the AI Force merely coordinates existing agencies operating under existing law, an executive order may provide a legitimate starting point. If it is supposed to compel disclosures from private companies, require incident reporting, regulate frontier models, impose penalties or exercise durable authority beyond powers Congress has already granted, Congress needs to define and empower it through legislation. Congress must also appropriate the money required to make the organization more than a title on an organizational chart.

If Washington is serious about an AI Force, the administration should send Congress proposed legislation defining its mission, limits, authority, funding and accountability. Congress should debate it, amend it and, if agreement can be reached, send it to the president for his signature. That is slower and less marketable than a Truth Social announcement. It is also how a constitutional government creates an institution intended to survive presidents, political cycles and court challenges.

There are hints that something may exist behind this latest announcement. The Information reported that a draft executive order creating an AI self-regulatory organization had circulated within the administration but recently stalled. Former AI and crypto czar David Sacks reportedly opposed creating such a body to monitor AI model developers.

We do not know whether that proposed organization and Trump’s AI Force are the same thing. He may be reviving the stalled proposal, rebranding some version of it or announcing something entirely different. Until the administration releases an order or a charter, anyone claiming to know is guessing.

But the report suggests that a real policy fight may be taking place behind the marketing.

The Government Already Has the Pieces

Washington does not need to build an AI Force from an empty lot.

The administration’s own AI Action Plan distributes AI responsibilities across NIST, the Department of Homeland Security, CISA, the Department of Defense, the Department of Energy, the intelligence community, the Office of Science and Technology Policy, the National Security Council and sector regulators.

The plan calls for frontier-model evaluations, testbeds, vulnerability sharing, incident response, security standards and research into interpretability and control. It says the government must monitor emerging and unforeseen AI risks with “constant vigilance.”

Those are not empty aspirations. They are many of the capabilities a serious AI Force would require. The problem is that they are divided across agencies with different missions, legal authorities and levels of technical expertise.

The closest thing to a technical core already exists inside NIST. The Center for AI Standards and Innovation, known as CAISI, develops AI security evaluation methods, works with developers and evaluators, examines national-security risks and helps establish standards and best practices. It looks like the natural technical arm of any credible AI Force.

But CAISI illustrates the gap between capability and force. Its arrangements with private AI developers are largely voluntary. It also reportedly needs substantially more resources. Before Washington creates another organization, it should explain whether the AI Force will fund and empower the one it already has, place an operational structure around it or simply duplicate it with a better name.

The best version would put an operational structure around capabilities the government already possesses. The worst would duplicate them and give the duplication a better name.

A Czar Is Not a Force

The promised appointment will offer the first meaningful clue.

David Sacks left the formal AI and crypto czar position in March after serving as a special government employee, although he continues to advise the administration. At that point, the White House reportedly did not plan to replace him. Trump has now revived the job.

Who gets it will tell us plenty. Will the new czar be a government official or an outside adviser? An industry advocate or an independent watchdog? A technical expert, a policy operator or a political messenger? Will the position have a staff, a budget and the ability to coordinate agencies? Or will the czar have a title, access to the president and little else?

An adviser can recommend. A committee can discuss. A regulator can compel. A force should be able to act.

There are at least four organizations hiding inside Trump’s label: An industry promotion council, a White House coordinating task force, a technical safety body or an enforcement organization. Those are not interchangeable missions.

Trump’s announcement tries to combine several of them. He wants the government to cherish AI, help it grow, watch over it and look for bad behavior. That creates an obvious conflict. The AI Force cannot simultaneously be AI’s cheerleader, coach, referee and police officer.

The administration needs to decide what job it wants this organization to perform.

Courts Are Not an Early-Warning System

Trump is not entirely wrong when he says existing criminal and civil law can deal with bad conduct involving AI.

Fraud does not become legal because an algorithm commits it. Discrimination does not become acceptable because a model made the decision. Negligence, unfair business practices, product liability and existing sector regulations do not vanish when artificial intelligence enters the process. Prosecutors, the Federal Trade Commission and other regulators already have authorities they can use.

The problem is timing.

Courts and prosecutors generally become involved after conduct has occurred, harm has been alleged and someone can be identified as responsible. That process is essential, but it is not the same as continuous technical oversight.

Who receives reports of serious AI near misses? Who can inspect model logs or test results? Who evaluates an autonomous system before it reaches critical infrastructure? Who coordinates a failure crossing several regulated industries? Who acts when a deployment appears dangerous but has not yet produced a crime or a plaintiff?

The criminal and civil justice systems can help determine who is responsible after an AI system causes damage. They are not the country’s AI early-warning system.

I have argued that a universal slowdown in AI development is unrealistic. Countries will not surrender an advantage because their competitors promise to do the same. Neither will companies. But rejecting an unenforceable slowdown does not require us to choose inaction until something goes badly wrong.

The practical alternative is active governance: Independent evaluation, incident reporting, technical monitoring, clear responsibility and human control. An AI Force could provide that machinery, but only if it has machinery.

What An Actual Force Would Look Like

A credible AI Force does not need to become an all-powerful national AI regulator. It does need a defined and limited mission.

It should coordinate agencies, establish severity levels for AI incidents and receive reports of serious failures and near misses. It should direct independent evaluations of high-risk systems, obtain relevant technical information under confidentiality protections and coordinate emergencies involving critical infrastructure. Violations should be referred to agencies possessing enforcement authority.

It should establish standards for federal AI, report publicly on risks and preparedness, and tell Congress where existing law leaves dangerous gaps.

The president can begin the work by coordinating agencies, setting federal procurement requirements and directing evaluation programs under existing authority. But if the AI Force is supposed to possess compulsory powers and survive this administration, Trump will need Congress. That is not an excuse for doing nothing. It is the constitutional path from announcement to institution.

Look for a written order, a defined mission, a leader, staff, participating agencies, a budget and a formal relationship with CAISI. Look for incident reporting, technical evaluation and clear enforcement or referral authority. Most of all, look for legislation.

Until some of those things appear, the AI Force remains a brand waiting for an institution.

The idea itself deserves a fair hearing. AI crosses national security, critical infrastructure, finance, health care, employment, communications and consumer protection. No single federal agency sees the entire picture. A focused coordinating body could fill a genuine gap.

There is also a contradiction worth watching. Trump continues to dismiss many AI warnings as a hoax. Yet reviving the AI czar position and announcing an AI Force suggest that his administration is responding to those warnings, even if he will not acknowledge doing so. The rhetoric says there is little to worry about. The institutional response suggests otherwise.

President Trump has already given the AI Force a powerful name. He has already launched the marketing campaign. Now he must tell us what it can see, what it can investigate and what it can do. More importantly, he must work with Congress to define its mission, establish its limits and give it the lawful authority and funding it needs to succeed within our constitutional system.

Until then, the obvious question remains: Where is the force in the AI Force?