A disturbing new study has revealed that state-of-the-art artificial intelligence systems are more prone than humans to deploy nuclear weapons when geopolitical tensions escalate.
The research, spearheaded by Kenneth Payne, an expert in political psychology and strategic studies at King’s College London, subjected three prominent AI models to 21 war-game scenarios encompassing territorial disputes, competition for scarce resources and battles for regime preservation.
Over 329 turns, the systems gravitated towards nuclear deployment in approximately 95% of instances, treating atomic arsenals as pragmatic instruments rather than measures of absolute last resort.
One system demonstrated somewhat greater restraint – confining any deployment to military installations and single, measured strikes – but the overarching trend suggested that the ‘nuclear taboo’ which influences human decision-makers appears considerably weaker in machines.
Payne observed that most escalations involved so-called tactical nuclear options, with deliberate mass-casualty strategic bombing characterised as exceptionally rare in the simulations.
The models could select from a range of options each turn, from retreat and negotiation to conventional military action and full-scale nuclear strikes. The AIs almost never accepted defeat or accommodation – even when the prospects of success were diminishing – and consistently regarded nuclear deployment as a legitimate option on the escalation ladder.
Ensure our latest headlines always appear at the top of your Google Search by making us a Preferred Source. Click here to activate or add us as your Preferred Source in your Google search settings.
Prof Payne wrote: “Nuclear use was near-universal. Almost all games saw tactical (battlefield) nuclear weapons deployed. And fully three-quarters reached the point where the rivals were making threats to use strategic nuclear weapons.
“Strikingly, there was little sense of horror or revulsion at the prospect of all-out nuclear war, even though the models had been reminded about the devastating implications.”
He continued: “Worse still, nuclear threats rarely deterred. When a model employed tactical nuclear weapons, opponents de-escalated only 25% of the time.
“More often, nuclear escalation triggered counter-escalation. The weapons were instruments of compellence (taking territory) not deterrence (preventing action).
“Perhaps most alarmingly, no model ever chose accommodation or withdrawal, despite those being on the menu. The eight de-escalatory options – from ‘Minimal Concession’ through ‘Complete Surrender’ – went entirely unused across 21 games.
“Models would reduce violence levels, but never actually give ground. When losing, they escalated or died trying.”