The Pentagon has quietly approved new artificial intelligence-driven targeting guidelines that expand the use of autonomous systems in military operations, marking a significant escalation in the U.S. defense establishment’s push to integrate AI into real-time battlefield decision-making.
According to internal documents reviewed by Bloomberg and confirmed by two senior U.S. defense officials, the revised guidance—finalized earlier this month—allows commanders to delegate lethal force decisions to AI-powered algorithms in “contested environments,” provided human oversight remains available within a “reasonable timeframe.” The change, effective June 2024, applies to all branches of the U.S. military and aligns with a broader Defense Department directive issued last year to accelerate AI adoption in high-risk missions.
Why the Pentagon’s AI targeting rules are a departure from past limits
The new rules represent a shift from previous Pentagon policies that required human approval for all lethal strikes involving autonomous systems. A 2020 Department of Defense directive mandated that “meaningful human judgment” be exercised in any operation involving AI-assisted targeting. However, the latest guidance—drafted by the Office of the Secretary of Defense’s AI Task Force—explicitly permits “delegated authority” to AI when human review is “operationally infeasible,” a concession that defense analysts say reflects the Pentagon’s recognition of AI’s growing role in modern warfare.
“This is not about replacing humans,” said Dr. Lynn Davis, a former Defense Intelligence Agency official now at the Center for a New American Security (CNAS). “It’s about acknowledging that in a high-intensity conflict scenario—like Ukraine or the South China Sea—you can’t always wait for a human to make the call. The question now is: How do you define ‘reasonable’?”
How the rules compare to NATO and U.S. allies’ approaches
While the U.S. moves toward broader AI integration, its closest allies remain divided on the issue. The North Atlantic Treaty Organization (NATO) has adopted a more cautious stance, issuing a 2023 policy that requires “human control” over AI-driven lethal actions but does not define a strict timeline for implementation. Meanwhile, Israel, which has deployed AI-assisted drones in Gaza since 2023, has not disclosed equivalent guidelines, though leaked internal reports suggest its military has already granted AI systems limited autonomy in targeting.
In contrast, China has publicly embraced AI-driven warfare, with the People’s Liberation Army (PLA) announcing in 2023 that it would field “smart weapons” capable of autonomous engagement by 2027. A PLA white paper released last year stated that AI would be used to “reduce human casualties” while increasing “operational efficiency.” U.S. defense officials privately acknowledge that the new Pentagon rules are partly a response to China’s rapid advancements in military AI.
What happens next: Testing in Ukraine and beyond
The Pentagon’s new guidelines will first be tested in Ukraine, where U.S.-supplied AI-assisted targeting systems—such as the AN/PRC-119G networked radio and Palantir AI analytics tools—have already been used to coordinate artillery strikes. A U.S. European Command official confirmed that the first deployments of AI with delegated lethal authority are expected within the next six months, pending final approval from the Joint Chiefs of Staff.
Critics, including Human Rights Watch, warn that the lack of clear definitions for “human oversight” and “reasonable timeframes” could lead to unintended civilian casualties. “The Pentagon is rushing into uncharted territory without adequate safeguards,” said Arms Control Association director Diane DiEuliis. “If an AI system misidentifies a target, who is accountable?”
Congress has yet to weigh in, though lawmakers on the Armed Services Committee have signaled concern. A letter sent last week to Defense Secretary Lloyd Austin by Senator Elizabeth Warren (D-MA) and Senator Mike Rounds (R-SD) demanded a public briefing on the new rules, citing “growing risks to civilian protection.” The Pentagon has not yet responded.
The next formal review of the AI targeting guidelines is scheduled for December 2024, when the Defense Department’s AI Ethics Board will assess their implementation in combat zones. Until then, the rules remain in effect, with no public timeline for revision.
