Helen Warrell, ft Investigative reporter
It’s July 2027, and China is on the verge of invading Taiwan. Autonomous drones with AI targeting capabilities are intended to overcome the island’s air defenses as energy supplies and key communications are cut off in order to cripple Amy-infiltrated cyberattacks. Meanwhile, a massive campaign implemented by an AI-powered pro-Chinese meme farm spreads across global social media, debunking Beijing’s act of aggression.
Such scenarios have brought dystopian horror to the debate about the use of AI in warfare. Military commanders hope for a digitally augmented force that is faster and more accurate than human-directed combat. But there are fears that as AI increasingly takes a central role, these same commanders will control a conflict that escalates too quickly and lacks moral or legal oversight. Former US Secretary of State Henry Kissinger spent his last years Warning About the coming catastrophe of AI-driven warfare.
Capturing and mitigating these threats is a military priority – some would call the “Oppenheimer moment” of our age. An emerging consensus in the West is that decisions surrounding the deployment of nuclear weapons should not be outsourced to AI. UN Secretary-General Antonio Guterres has gone further, calling for an outright ban on more autonomous lethal weapons systems. It is important that regulation keeps pace with evolving technology. But in the sci-fi-fueled excitement, it’s easy to lose track of what’s actually possible. As researchers at Harvard’s Belfer Center point out, AI optimists often underestimate the challenges Fielding a fully autonomous weapon system. It’s entirely possible that the AI’s abilities are being overpowered in combat.
Anthony King, director of the Strategy and Security Institute at the University of Exeter and a leading proponent of this argument, suggests that rather than replacing humans, AI will be used to improve military intelligence. Even if the character of war is changing and remote technology is improving weapon systems, he insists, “complete automation of war itself is just an illusion.”
Of the three current military use cases of AI, none involve full autonomy. It is being developed for planning and logistics, cyberwarfare (sabotage, espionage, hacking, and information operations); and, most controversially, for targeting weapons, an application already used in the Ukraine and Gaza Strips. The AI-assisted decision support system, known as Lavender, has helped identify nearly 37,000 potential human targets inside Gaza.
FT/MIT Technology Review | Adobe stock
There is clearly a risk that the Lavender database replicates the biases of the data it is trained on. But military personnel are also biased. An Israeli intelligence officer who used lavender claimed to have more confidence in A’s fairness.Statistical procedures“Compared to the Sorrowful Soldier.
Tech optimists designing AI weapons also deny that specific new controls are needed to control their abilities. Keith Dear, a former British military officer who now runs the strategic forecasting company CassAI, says the current rules are more than adequate: “You make sure there’s nothing in the training data that corrupts the system… when you’re confident you deploy it.