The new AI arms race changing the war in Ukraine (www.bbc.com)

Ukrainian forces recently intercepted a Russian drone that, unlike conventional models, used onboard AI to find and strike targets autonomously without emitting radio signals—making it resistant to jamming. Both Kyiv and Moscow are already deploying AI across the battlefield: Ukraine reports more than 50,000 frontline video streams a month that AI parses to identify targets and map threats, while drones increasingly execute final autonomous approach phases, lock on targets, drop ordnance, assess damage and return. Developers envision systems that a soldier could trigger from a smartphone and that could operate as interceptors against long‑range attack drones like Shaheds. Ukraine claims partial implementations are fielded and thousands of automated systems could be deployed by 2026. Those advances promise faster decision cycles, denser sensor fusion and force‑multiplying swarm tactics, but they also deepen ethical, legal and tactical dilemmas. Automated weapons risk misidentifying combatants or civilians—manufacturers like DevDroid deliberately disable automatic firing to reduce friendly‑fire risks—and there are unresolved questions about compliance with international humanitarian law and surrender protocols. Jamming, kinetic interception and traditional air defenses may be ineffective against low‑signature, autonomous swarms, heightening the urgency of countermeasures and global norms. President Zelensky and others warn this fuels “the most destructive arms race,” prompting calls for international rules on AI in weapons akin to non‑proliferation regimes.