Future AI weapons such as drones should be programmed with moral guidelines, according to former UK spy chief David Omand, who has changed his mind on unmanned weapons systems after more than a decade.
Omand’s Change of Heart
David Omand, former head of GCHQ, told the Guardian that he now believes autonomous drones can comply with international humanitarian law. He had previously concluded that such compliance was impossible. Omand said he was swayed by the speeding-up of modern warfare, including the deployment of drones and hypersonic missiles, and the emergence of generative AI, which offers a potentially ethical way for drones to operate under morally compliant systems.
Moral Frameworks for Machines
Omand argued that AI technology can now weigh factors involved in human drone operators’ targeting decisions, such as target legitimacy, civilian casualties, and correct identification. This is not inventing new ethics but putting existing military ethics into a form deployable by machines. He called for urgent work to ensure future AI-powered weapons have a moral component built in.
His intervention comes as armed forces minister Al Carns said there might be circumstances where machines make their own targeting decisions, requiring the ability to take the human out of the loop. Meanwhile, the US is heavily investing in AI-powered warfare, with a $54bn allocation for autonomous systems in its 2027 budget.
Human Oversight: ‘On the Loop’
Omand believes that as warfare becomes more technologically sophisticated, humans will inevitably need to be ‘on the loop’ rather than ‘in the loop’ with AI systems. ‘On the loop’ means a human supervises the system but does not authorise every combat action, with generative AI enabling machines to operate within a human-designed moral framework. These frameworks must be adjustable from operation to operation.
Moving to an ‘on the loop’ setup is a physical and operational inevitability, Omand said, because in high-speed warfare, there is no time for human decision-making in individual combat actions. The question now is how to ensure autonomous weapons adhere to international law.
Adaptive Moral Control Layer
Omand proposes an ‘adaptive moral control layer’ where a human sets parameters such as the expected proportion of civilians around a target before a mission. This formalises moral authority. He argues that such a system could be ethically superior to human decision-making, potentially reducing collateral damage in fast-paced warfare.
However, Chris Cole, director of Drone Wars UK, called Omand’s stance ‘as nonsensical as it is dangerous’, stating that AI is simply not capable of making judgments, only processing data, and lacks the ability to distinguish civilians from combatants or judge proportionality.
