Recent reports detail a daring Ukrainian operation, dubbed “Spider’s Web,” involving drone strikes on Russian air bases. The attack has sparked discussions about the evolving nature of warfare and the potential role of artificial General Intelligence (AGI).

On June 1,Ukraine reportedly “targeted” several Russian air bases using first-person view (FPV) drones,remotely operated aerial vehicles. According to “reports” Ukraine employed machine-learning algorithms to guide the drones.

The “Spider’s Web” attack demonstrated the hardware capabilities of modern warfare and provided a glimpse into what the future of warfare could look like, especially as companies and governments develop “Artificial General Intelligence” (AGI).

The “Spider’s Web” Operation

The Security Service of Ukraine’s (SBU) operation reached targets up to 8,000 kilometers (nearly 5,000 miles) from the frontlines. Analysts are now “wonder[ing]” whether anywhere is truly beyond the reach of FPV drones.

Some “reports” suggest dozens of strategic bombers were destroyed or disabled by 117 FPV drones, though Moscow countered that only a handful of planes were struck. Western “assessments” put the figure at no lower than 10.

While the scale of the attack was remarkable, its logistical brilliance is noteworthy. The operation followed a Russian drone attack at the end of May “involving” almost 500 unmanned aerial vehicles.

According to the “BBC,” parts were smuggled into the country and the drones were assembled. Vasyl Maliuk, the head of the SBU, said that they were then loaded onto lorries with secret compartments and driven by unsuspecting Russian couriers to locations near air bases. When the shipments reached their destination, roofs on the lorries retracted to reveal the hidden hardware, and the drones took off.

“Spider’s Web” depended on logistics, deception, and coordination.

The attack confirms that expendable drones are the weapon of the 21st century and serves as an example of how AGI will work as a warfighting tool and how humans will work alongside AGI.

The Rise of AGI-Powered Warfare

AGI-fueled warcraft is on the horizon. Over the past two years, the AI industry has “increasingly invested” in military applications of AI.

Frontier labs are embedding themselves into the national security state. For instance,in June 2024 OpenAI “appointed” retired U.S. Army general Paul Nakasone to its Board of Directors. In December 2024, the AI-giant announced it had “partnered” with defense military technology outfit Anduril to develop drone defense systems. And Google “scoped out” “national security imperatives for the AI era” earlier this year.

The technology sectors’ allusions to national security and AI have a shape-shifting quality. It’s not always clear whether someone is referring to defensive or offensive AI capabilities.

The idea that a sufficiently capable AGI system might eventually pilot drones is “already on the minds” of military planners, but Ukraine’s strike on Russia gives us a more specific picture of what to expect.

“Spider’s Web” had been “in the making” for eighteen months. During this time,150 small attack drones and 300 explosive devices were smuggled into Russia to stage the attack.

The SBU likely engaged in piecemeal smuggling to avoid detection, possibly bringing components across borders, using front companies, or bribing officials to pass through checkpoints.

The final drones were packed into special mobile containers that looked inconspicuous from the outside.

According to “reports” the drivers of the lorries all told a similar story. A businessman approached them to pick up what seemed to be wooden cabins and deliver them to various locations around Russia. They agreed and thoght little of it.

Once the trucks were in position, the strike was launched. At the predetermined moment, each container’s roof panels were remotely opened to release a swarm of drones (likely piloted remotely by “piggybacking” on Russian telecommunications networks).

AGI’s Potential Impact on Future attacks

The “Spider’s Web” operation offers a glimpse into how AGI could supercharge similar attacks in the future.

AGI could analyze transportation routes to find the safest, fastest, and least conspicuous way to move cargo. It could plan truck routes that avoid busy checkpoints, choose transit times when border guards are understaffed, and even account for satellite overpasses or drone surveillance.

Such a system could coordinate multimodal logistics with timing that no human team could match. It could crunch traffic patterns, rail schedules, and weather data to find the perfect moment for an attack.

This hypothetical warfighting AGI could automatically generate corporate entities complete with registration documents, tax records, and websites to serve as cover. It could forge driver’s licenses, passports, and employee IDs that pass automated verification.

Aside from paperwork, an AGI could manage a whole suite of deception technologies. For example, AGI could emit fake GPS signals to confuse satellite tracking or hacking into a facility’s CCTV feed to loop old footage while operatives move equipment. When it’s time to strike, AGI could guide each drone to its target as part of a single unified swarm, optimized to prevent collisions and spaced to maximize coverage.

AGI may even make it possible to monitor the electronic warfare environment and switch frequencies if it senses jamming on the current channel. If an air defense system starts tracking the swarm, the AGI might command all drones to disperse or drop to terrain-hugging altitude to increase their odds of survival.

As soon as the destination is in range, AGI could help drones autonomously recognize target types and aim for the most damaging impact points.

Limitations and dangers

These are still predictions about what AGI might potentially be capable of in the future, and there will likely be limitations.

Precision hand-work