Pentagon AI Weapons Fall Short of Own Guidelines

Pentagon AI Weapons Fall Short of Own Guidelines, Watchdog Finds

A Government Accountability Office report found that six of 10 Pentagon officials surveyed identified operational risks in AI projects despite the agency's stated ethical guidelines.

The Political Desk

Gokhshtein Media

October 24, 2026

3 min read

Officials say the Pentagon's artificial intelligence-powered weapons and projects don't always follow the agency's internal guidelines for AI, according to a congressional watchdog report published Oct. 14, 2026. The U.S. Government Accountability Office found that some officials involved with these tools identified operational risks.

The GAO surveyed 10 officials working on 10 specific AI weapons and tools. These projects use AI for automatically identifying battlefield targets, processing surveillance data and enhancing tactical decision-making.

Six of the 10 officials stated their teams had "identified risk(s)" when considering using these projects "in an operational environment." These 10 projects are part of a larger portfolio of more than 300 AI initiatives currently underway within the Pentagon.

Defense Secretary Pete Hegseth has consistently pushed for rapid autonomous innovation within the Pentagon, dismissing previous warnings about AI tools. In January 2026, the Pentagon issued a memo declaring its intention to become an "AI-first" warfighting force.

Last month, Secretary Hegseth announced the establishment of a new warfare command tasked with advancing autonomous weapon development, including drones.

In 2021, the Pentagon issued a memo outlining six basic tenets designed to ensure responsible development of AI weapons. These guidelines emphasize Pentagon-wide oversight, verification of the technology and real-time monitoring. They also stress cooperation with allies to foster global norms based on shared values.

Alissa Czyz, who directed the GAO review, explained that the 2021 memo was not a formal department instruction but was understood as an important directive for those involved in building AI tools within the Pentagon. "It's something that the department has signaled that should be implemented," Czyz said.

The GAO report indicated that most personnel involved in the 10 reviewed projects had taken at least initial steps to follow the six tenets. Czyz also noted that some of these projects were still in early stages of development.

AI researchers and international organizations have long advocated for strict regulation of AI in weapons, arguing that such tools could become unpredictable, operate with minimal human control, endanger soldiers and civilians and increase the risk of rapid war escalation.

Last month, Secretary Hegseth announced the establishment of a new warfare command tasked with advancing autonomous weapon development, including drones.