Former GCHQ Chief Reverses Stance on Autonomous AI Weapons

A former head of Britain's signals intelligence agency has reversed his decade-old opposition to autonomous weapons systems, arguing that artificial intelligence can now enable drones to make ethically sound combat decisions.

David Omand, who led GCHQ from 1996 to 1997, told The Guardian he now believes AI technology can create moral frameworks for unmanned weapons that distinguish between combatants and civilians—a capability he previously deemed impossible.

Omand chaired a 2014 University of Birmingham commission that expressed serious doubts about whether autonomous systems could safely identify civilians or exercise the proportionality required under international humanitarian law. His reversal comes as military forces worldwide accelerate AI integration into combat operations.

The case for machine morality

Omand points to two developments that changed his assessment: the accelerating tempo of modern warfare and advances in generative AI. Weapons including drones and hypersonic missiles now operate at speeds that outpace human decision-making, he argues, making some level of machine autonomy operationally inevitable.

The former intelligence chief proposes what he calls an "adaptive moral control layer"—a system where human operators set moral parameters before missions, which AI then applies during combat. These parameters would address factors like target legitimacy, expected civilian presence, and proportionality of force.

"It could actually work, whereas relying on humans in a very fast-moving seconds matter for warfare is probably going to lead to far worse results in terms of collateral damage," Omand said.

He distinguishes between humans being "in the loop"—authorizing every action—and "on the loop," where they supervise systems without approving individual strikes. The latter approach, he contends, could produce ethically superior outcomes to split-second human judgments under pressure.

Military momentum builds

Omand's comments align with shifting military policy. UK armed forces minister Al Carns recently stated there may be circumstances requiring the ability to "take the human out of the loop when required."

The United States has committed $54 billion in its 2027 budget to autonomous and remotely operated systems across all domains, including a "drone dominance" program. AI from companies including Palantir and Anthropic has already been deployed in recent operations to compress targeting timelines, according to The Guardian's reporting.

Current UK defense policy requires "context-appropriate human involvement" in weapons that select and attack targets, which can include control through operational parameters rather than individual strike authorization.

Why it matters

The debate over autonomous weapons sits at the intersection of military necessity, technological capability, and international law. As AI systems become more sophisticated and warfare accelerates, militaries face pressure to delegate more decisions to machines. Omand's reversal—from a respected intelligence veteran who previously opposed such systems—signals growing acceptance of autonomous weapons within defense establishments. Whether AI can truly encode the moral reasoning required by laws of war remains contested, with implications for civilian protection in future conflicts.

Criticism and concerns

Chris Cole, director of Drone Wars UK, rejected Omand's position as "as nonsensical as it is dangerous." Cole argued that AI cannot make genuine moral judgments, only process data, and lacks the capacity to distinguish civilians from combatants or assess proportionality.

Omand, now a visiting professor at King's College London's war studies department, has held board positions at defense companies and advises Paladin Capital, a cybersecurity investment firm.

The details were first reported by The Guardian.