The Trump administration’s aggressive push to streamline federal agencies continues to reverberate, with the latest casualty being a critical oversight body within the Department of Defense. As reported by MIT Technology Review, Secretary of Defense Pete Hegseth announced on May 28 a drastic cut to the Office of the Director of Operational Test and Evaluation (DOTE)—a department tasked with independently testing the safety and performance of weapons and AI systems before they reach the battlefield.
The decision reduces the office’s staff from 94 to about 45 and removes its director, with only a week granted for implementation. This move represents an unprecedented overhaul of an agency that has operated largely unchallenged for four decades. Founded in the 1980s to combat the Pentagon’s history of deploying underperforming technologies, the DOTE has long served as a final checkpoint before military systems are deployed at scale.
Missy Cummings, a former Navy fighter pilot and now a professor at George Mason University, described the office as “the last gate before a technology gets to the field.” While small-scale tests can bypass DOTE, any broader deployment must pass through its rigorous evaluations—evaluations that have historically reduced waste, fraud, and abuse through independent scrutiny of defense contractors’ claims.
The timing of this gutting raises particular concern, given the Pentagon’s growing interest in artificial intelligence. With AI now being incorporated into everything from autonomous drones to battlefield decision-making tools, the need for careful and transparent testing is more pressing than ever. The tech-military pipeline is booming, with defense startups like Anduril securing multibillion-dollar contracts and OpenAI expressing increasing comfort with military partnerships.
Hegseth argues the cuts will streamline operations and save $300 million. But experts warn that this efficiency could come at a cost. “The firings in DOTE send a clear message that all perceived obstacles for companies favored by Trump are going to be removed,” said Cummings. Mark Cancian, a senior advisor at the Center for Strategic and International Studies and a former Pentagon official, echoed that concern. While some in the military have criticized DOTE for slowing down deployment, Cancian emphasized that it frequently catches issues that could prove critical in combat.
This is especially true when dealing with technologies like generative AI, which are inherently unpredictable. Systems that function well in controlled lab conditions often behave differently in complex, real-world environments. DOTE’s evaluations help identify these risks before they become liabilities—but now, with half the personnel, its ability to do so is significantly diminished.
Though it’s unclear whether companies like Anduril, Anthropic, or OpenAI lobbied for the cuts, their growing influence and proximity to military leadership raise questions. None responded directly to inquiries about their stance on the restructuring, but the shift undeniably benefits firms racing to roll out new AI systems for defense.
Ultimately, the move underscores a growing tension between rapid innovation and responsible oversight. While the Pentagon’s enthusiasm for AI is not new, the current wave of ambition—with a particular focus on large language models—introduces a new layer of complexity and risk. Cutting the one office designed to vet those risks doesn’t seem like a win for national security, no matter how efficiently the technology gets deployed.
