The next phase of the drone war in Ukraine may be decided less by who has the most drones and more by who can shorten the time between detection and response.
That is where artificial intelligence is entering the fight.
Russia has expanded its use of faster, jet-powered drones that are harder for Ukrainian air-defense teams to intercept. Associated Press reporting on October 1 described Ukrainian manufacturers racing to build faster interceptors and using AI to help guide them toward incoming threats.
This is a legitimate defensive use of AI. It is also a warning about how quickly automation can move from experimental technology into high-stakes security decisions.
The problem is speed
Traditional air defense depends on detection, classification, communication and interception. Each step takes time.
Faster drones compress that timeline. A human operator may have only moments to identify a threat and decide how to respond. AI can help process sensor information faster, prioritize tracks and improve the odds that an interceptor reaches the right target.
In that environment, automation is not simply about efficiency. It becomes part of survival.
But the same advantage creates a new risk
The faster a system operates, the less time humans have to question it.
An AI-supported interceptor can reduce workload and improve reaction time. But battlefield environments are noisy. Sensors are imperfect. Electronic warfare can distort information. Civilian aircraft, friendly systems and decoys can complicate classification.
If the system becomes too confident, too opaque or too fast for meaningful oversight, the advantage of speed can become the risk of error.
The challenge is not whether AI should be used. It is deciding where human judgment still has to remain decisive.
Ukraine is also showing how quickly capability spreads
The technologies involved are not limited to large defense contractors. Commercial components, software, machine vision and autonomous navigation are increasingly available to smaller firms and, eventually, to non-state actors.
That is the dual-use problem.
The same advances that help a country defend a city can eventually help armed groups improve surveillance, targeting, propaganda, logistics or drone operations. That does not mean every new model or component becomes a weapon. It means the barrier between civilian innovation and military adaptation is getting thinner.
Security policy has to catch up with that reality without pretending innovation can simply be stopped.
The real lesson is not “AI wins wars”
Technology rarely works that cleanly.
AI can improve detection, help manage large amounts of data and support faster defensive decisions. It can also be spoofed, degraded, misused or overtrusted. An adversary adapts. The system adapts back. The advantage is temporary.
What matters is whether institutions can integrate the technology without surrendering judgment to it.
That requires testing under real conditions, visible confidence levels, clear rules for human intervention and a willingness to slow the system when uncertainty is too high.
The conclusion
Ukraine’s air-defense challenge is becoming a real-world test of AI under pressure.
If AI helps defenders react faster, it can save lives. But speed should not become the only measure of success.
The stronger system will be the one that combines automation with disciplined human control. In security, the goal is not simply to make machines decide faster. It is to make people better able to decide when the machine may be wrong.
Ahmad Shah Mohibi
Founder, Rise to Peace