General’s AI Warning Rattles Pentagon

A hand interacting with a digital interface displaying AI technology

A retired three-star general who helped build the Army’s future force is now warning that artificial intelligence could reshape war faster than our safeguards can keep up.

Story Snapshot

A retired Army lieutenant general who led ground robotics and autonomy efforts is now deep inside the defense AI industry.
The War Department is racing toward an “AI‑first” military with 30‑day deployment timelines and wartime‑speed experimentation.
Critics warn that rushed AI adoption can sideline human judgment, threaten civil liberties, and hide failures from the public.
Conservatives must insist AI strengthens American power without feeding a permanent war‑tech bureaucracy or eroding constitutional rights.

A general who helped build the AI future is now on the outside looking in

Retired Army Lieutenant General Ross Coffman spent the final years of his 35‑year career at the center of military modernization.^[7] He served as Deputy Commanding General of Army Futures Command, the second‑highest role in the organization responsible for reshaping the force, with a heavy focus on ground robotics and autonomy.^[1] Before that, he directed the Next Generation Combat Vehicle team, standing up the Army’s ground autonomy efforts that tied robots, vehicles, and weapons into one data network.^[1]

After retirement, Coffman did not walk away from technology. He became president of Forward Edge AI, a defense‑oriented artificial intelligence company, and joined autonomous‑trucking and defense firm Kodiak AI as a strategic advisor.^[1]^[7] Kodiak says he will guide how to deploy its autonomous systems for national security and chair a new Defense Advisory Council, giving him a direct line into how military AI is built and sold today.^[1] He also sits on boards and speaks on how to lead in the age of AI.^[3]

Washington is pushing “AI‑first” war at wartime speed

While Coffman and other veterans raise questions, the War Department under President Trump has adopted a blunt motto for the future: speed wins. Its Artificial Intelligence Strategy orders the department to become an “AI‑first” warfighting force and directs leaders to accelerate “America’s military AI dominance.” A related strategy memo tells officials to deploy the latest commercial AI models into military systems within 30 days of public release, treating software like ammunition that must reach the front fast.

An AI Acceleration Strategy press release says the department is taking a wartime approach, tearing down bureaucratic blockers and driving seven “pace‑setting projects” across warfighting, intelligence, and support missions. One project, called “Agent Network,” aims to unleash AI agents for battle management and decision support, from campaign plans down to trigger‑pulling kill chains. Another effort, “GenAI.mil,” promises frontier generative models for personnel across the force, making advanced AI tools as common as email for many jobs.

Supporters promise control, but proof of real safeguards is thin

Policy papers from the War Department and the White House insist that rapid adoption and safety can go together. America’s AI Action Plan talks about secure computing environments, testing standards, and risk checks for national security uses, all meant to keep powerful systems under control while government moves fast. An Army University Press article argues the Army must urgently modernize decision‑making with AI, but says human commanders should stay in charge through careful testing and clear rules.

Yet these documents are plans and strategies, not hard evidence. They do not provide public test results showing AI systems have performed safely in real missions or high‑stress exercises. A War Department AI strategy memo calls for “robust experimentation” and metrics, but it does not share examples where AI was stopped or fixed after failing in the field. For now, citizens are asked to trust that internal processes and contractors are getting it right, without seeing the failures or lessons that would prove it.

Warnings: lost human judgment, secret data grabs, and tech dependence

Outside watchdogs see a different danger: a rush to wire warfighting to AI without enough transparency, human control, or respect for civil liberties. The Brennan Center for Justice explains that fast adoption threatens to displace human expertise in life‑and‑death decisions, putting both troops and civilians at risk when algorithms are wrong or biased. The report also warns that military AI often depends on bulk data about Americans, bought from commercial brokers with little judicial oversight, putting privacy and the Fourth Amendment under pressure.

The same report shows how the Pentagon’s growing dependence on private tech firms can weaken government control over the code steering critical systems. Companies own the software, the data, and sometimes even the cloud infrastructure. That makes it harder for elected leaders to fully inspect or challenge tools that shape targeting, surveillance, and intelligence. Congress has the power to set strong rules but has done little on military AI, and agencies can even waive some safeguards if they claim oversight slows operations. For conservatives, that should raise red flags about unaccountable power.

What this means for conservative patriots

For Trump supporters who back a strong military, the goal is clear: keep America’s edge but do not hand a blank check to the war‑tech machine. Coffman’s career shows that serious warriors can push new tools and still worry about how they are used. He helped field robots and networks to give our troops an unfair advantage, yet his industry roles and the wider debate highlight the need for honest oversight.^[1]^[3] Power must stay with human commanders who answer to the Constitution, not with black‑box code or unaccountable contractors.

That means demanding three things from Washington. First, real transparency on where AI is used in targeting, surveillance, and command systems, with regular reports to Congress and the public when possible. Second, clear laws that keep final lethal decisions in human hands and protect American citizens from dragnet data grabs and mass tracking. Third, strong audits of defense AI contracts, so taxpayer money builds tools that win wars rather than feed a permanent, unreviewed bureaucracy. Speed matters in war, but for a free republic, control matters more.