BEN MORGAN – A tactical revolution in Ukraine, but what next?
Ukraine’s war against Russia is rapidly becoming the world’s first large-scale AI-assisted drone war. In this guest blog, military analyst Ben Morgan examines how drone swarms, robotic logistics and AI-guided battlefield systems are reshaping modern combat — and asks what happens when machines begin taking ground instead of soldiers.
In recent months, we can see more evidence of the tactical revolution Ukraine’s drones are delivering. Ukraine’s bombing campaign against Russia’s oil and defence industries continues, providing no respite for the Russian economy. And, on 25 April 2026, Ukrainian drones destroyed Russian fighter jets at Shagol airfield in the Chelyabinsk region – 1,700 km from the border. A new record for a long-range drone attack.[i]
Meanwhile, on the ground, Ukraine’s drone-based ‘kill web’ has stopped Russia’s summer offensive. In fact, the Institute for the Study of War reports that “Russian forces in April 2026 suffered a net loss of territory controlled in the Ukrainian theater for the first time since Ukraine’s August 2024 incursion into Kursk Oblast.”[ii] Since November 2025, Russia’s monthly rate of advance has slowed down – a trend reinforced by this report. Wet spring weather contributes to slower Russian progress but there is evidence that the slow rate of advance is influenced more by tactical changes on the battlefield.
Specifically, Ukraine making greater use of ground and air drones in combined arms operations, including President Zelensky’s announcement that “for the first time in the history of this war, an enemy position was taken exclusively by unmanned platforms — ground systems and drones.“[iii] Zelensky’s announcement is further evidence that Ukraine continues to expand its use of ground drones in both offensive and defensive operations over the last twelve months.[iv]
This trend is not new; for example, in July 2025 Ukraine’s 3rd Assault Brigade, fighting near Kharkiv, reported that “Positions that had previously withstood two assaults by adjacent units were retaken by the 3rd Assault Brigade’s robots thanks to well-planned offensive actions. The cleared fortifications and a forest belt have been secured by our forces.”[v] A small but important incident indicating how robotic capabilities are evolving as ground drones closed with, and captured an enemy position. Two operations that provide insight into the future and raise questions about – What comes next in the automation of war?
How are drones changing warfare in Ukraine?
Already, we can see glimpses of the future as drones move from surveillance, target acquisition and individual attack roles into more coordinated offensive roles. Ukraine and Russia are both using swarms of piloted First Person View (FPV) attack drones to produce battlefield ‘effects’ similar to traditional artillery. For instance, ‘immediate neutralisation’ missions that replace artillery or mortars with drone strikes to support infantry to win the firefight and establish fire supremacy in contact. Drones are also used to stop offensive operations, swarms of FPVs descending on assaulting enemy forces providing ‘defensive fire’ to stop an attack. And, even drone ‘creeping barrages,’ FPV drones engaging enemy positions ahead of an advancing force in the same way that artillery has since the First World War.[vi]
Notably, the use of ground drones to provide logistics support for operations is also evolving rapidly but is under-reported. Ukraine claims that a large percentage of frontline logistic support is now provided by uncrewed vehicles that bring ammunition, food and medical supplies to frontline troops, and bring casualties from the frontline back for medical treatment.[vii] Dev.UA news service reports that “In the Azov brigade, most of the transportation to the front line has also been given to robotic complexes: they deliver 20-25 tons of property every month. A similar practice exists in the Airborne Assault Forces: one of the brigades transports 10-12 tons per month, which is approximately 80% of all supplies to the front line.”
Now though drones are evolving from being a support element within a human-centric combined arms team towards conducting independent fire and movement. Zelensky’s announcement is another indication of Ukraine’s experimentation with teams of drones working together to engage an enemy position. Ground drones performing ‘fire and movement’ to manoeuvre forward and capture a position. A recent Lowry Institute article describes how “Ukraine is also experimenting with combining robotic systems rather than using them separately. Kyiv recently showed a UGV designed to carry FPV drones to a concealed launch point and release them from the vehicle itself, which would allow ambushes from unexpected positions while keeping human operators farther from the front.” [viii]
However, it is important not to over-hype these capabilities at this stage because unlike air or sea drones, ground drones must contend with the complexity of traversing terrain. This requires skilled operators, or very capable AI. Recently, Jacob Parakilis writing for the Rand Corporation,opined that “When it comes to ground robots taking on infantry combat, there are a set of serious obstacles. The first is, quite literally, obstacles.”[ix] Parakilis also points out that even with rapid developments in AI capabilities – “Autonomous driving is a massive challenge. Residents of London may have seen Waymo autonomous cabs in recent weeks, moving through the city’s streets ahead of their public rollout. But following traffic laws and (more-or-less) consistent road markings is still a huge and complex task. Navigating a battlefield in a complex 3D environment is at least as complex, requiring a huge amount of processing power.”
Echoing Rand Corporation’s assessment, Defence One reported that John Hardie, from the Foundation for the Defence of Democracies discussed the difficulties ground drones face navigating across country with reporters in February. Hardie stated that “My understanding is that they’ve experimented with autonomous navigation, but it’s especially difficult with [unmanned ground vehicles] for that to be reliable. So I don’t think they’re there yet.”[x]
Additionally, maintenance and sustainment are still considerations for drones. Drones require technical support, weapons must be reloaded, batteries replaced or recharged, and regular maintenance undertaken.
Therefore, the emerging picture of the current state of drone integration into Russian and Ukrainian forces can be summarised, as follows:
- FPV drones are supplementing and sometimes replacing artillery as a tool to impose indirect fire ‘effects’ on the enemy. ‘Neutralising, suppressing and destroying’ enemy forces. By clustering drones and operators in larger centralised units both sides achieve economies of scale for maintenance, and guarantee ‘mass’ when FPVs are deployed to deliver these ‘effects’ on the enemy.
- Increasingly, ground drones are being used to provide logistics support and bridge the 5-20km deep ‘kill web’ that arcs across the frontline. Many, small drones, moving quietly backwards and forwards through this zone attracting less attention than trucks. CNN reports that “Land drones are much harder to spot and intercept than larger military vehicles. Compared with their aerial counterparts, they can operate in all weather conditions and carry much larger payloads.”[xi]
- Ground combat drones are proving themselves in defensive operations with examples of weapon carrying drones defending positions for many days with minimal maintenance.[xii]
- And, recent reporting indicates that air and ground combat drones are starting to be used together in tactical teams to assault enemy positions.
It is clear that drone use is constantly evolving, and that we are now on the verge of AI playing a more important role in how drones are controlled so it is important to pause and consider how AI will change the use of drones in land combat.
What happens when AI starts making battlefield decisions?
When discussing the impact of AI on the deployment of drones in battle a key consideration is the difficulty of navigating ground drones across country. Unlike in the air, or at sea, the complexity of the landscape means that it will be some time before ground drones can move autonomously – let alone make tactical decisions like; taking cover, outflanking a position or choosing the most concealed route to an objective.
Instead, AI’s short-term impact on battlefield drone technology is more likely to be in the air where navigation is simpler. AI is very good at simplifying interactions between humans and machines, it can receive an instruction then fly a drone to perform that task. For instance, circling an area and monitoring activity or terminal guidance of a munition to a target. The machine taking over managing the inflight course corrections required to achieve the desired result, simplifying the operator’s task.
Already, AI is employed for terminal guidance; or flying a drone to its target. This idea is not new and existing missiles like Javelin already simplify the human-machine interface in a similar manner. Javelin’s technology removes the complexity of guiding the missile to its target. The soldier firing it can simply ‘fire and forget’ knowing that the missile will fly to the target without further input, even if the target is moving. The notable difference between legacy systems and new drones is cost; each Javelin missile costs hundreds of thousands of dollars while the cost of an FPV drone is measured in thousands of dollars.
Another example of AI simplifying the human-machine interface relates to AI technology enabling drones to ‘swarm’ together to deliver synchronised surveillance or attacks. Recently CBS interviewed an American army officer studying the impact of drones and AI, Captain Ronan Sefton who works on the ‘Ukraine Lessons Learned Task Force.’ Sefton described the use of AI to assist swarming drones, stating “It’s debated often as to what [swarm technology] actually means. But it’s really just a lot of drones working together at one time. And you’re taking away a cognitive load from, say, a pilot. And one person might be able to control many drones.”[xiii]Essentially, AI could enable the effort of one human operator to be multiplied so that instead of requiring 100 human operators to pilot a swarm, only one operator is required. AI ensuring the swarm stays in formation and that individual drones perform tasks allocated to them.
AI can also be trained to operate drones with weapon systems that monitor engagement areas and respond autonomously to threats. For example, using a ground drone to monitor an area that automatically engages enemy forces that enter the specified engagement area. This is a simple application of existing AI, the legal or ethical risks of which can be mitigated by training (the AI) and by having robust rules of engagement.
Conclusion – An ethical discussion about drones is fast approaching
In conclusion, this tactical revolution is not without risk because as AI becomes more capable the potential for fully autonomous ground drones becomes a reality. Machines making decisions about inflicting violence on human beings is an uncomfortable situation that requires ethical consideration because autonomous robotic soldiers, able to kill without emotion are the stuff of science fiction nightmares.
However, any discussion about the use of drones in ground combat is an emotive topic because of humanity’s relationship with war, ground and violence. In war, possession of ground is the key metric of tactical success and the side that occupies, and dominates the landscape wins the battle. The industrialisation of war has produced artillery and aircraft that can bombard enemy positions but bombs and shells do not win land battles. Instead, the occupation of ground by human soldiers remains the fundamental condition of victory. Therefore, the ethical considerations surrounding combat remain very human-centric and there is considerable concern about machines making decisions that result in violence.
Currently, the historic paradigm is being challenged because ground drones start to become a substitute for human soldiers. But, in the foreseeable future the application of AI at the frontline is unlikely to produce completely autonomous drones. Instead, its impact is more likely to be reducing the cognitive load on operators; allowing them to do more by leading swarms of drones, or by assigning tasks to drones. Battlefield decisions about violence still governed by the drone’s operator. However, the battlefield is constantly evolving and the requirement to closely examine the ethical deployment of AI on the battlefield in drones may be closer than we think.
Ben Morgan is TDBs Military Blogger. a defence and security analyst specialising in modern warfare, military adaptation, and operational-level conflict analysis. He posts at Substack. If you like this content and want to support it you can ‘Buy me a Coffee’ here – buymeacoffee.com/benmorgan
[i] https://united24media.com/war-
[ii] https://understandingwar.org/
[iii] https://www.businessinsider.
[iv] https://kyivindependent.com/
[v] https://www.pravda.com.ua/eng/
Trench Artand
[vii] https://dev.ua/en/news/nrk-
[viii] https://www.lowyinstitute.org/
[ix] https://www.rand.org/pubs/
[x] https://www.defenseone.com/
[xi] https://edition.cnn.com/2026/
[xii] https://edition.cnn.com/2026/
[xiii] https://www.cbsnews.com/news/
There was a scary science-fiction short story whose name I have forgotten from sometime in the 70s I think, where autonomous drones became self replicating, stopped fighting each other for whichever side they were on, and started killing off human beings. One of them look like a child carrying a teddy bear. The story ends with the protagonist taking off on the last rocket 1st to go to the moon, but unfortunately one of the robots has sneaked on board. The implication being that we are all fucked.
Obviously we aren’t at this level yet but people do tend to underestimate whatever advances in science may appear in the near future. I have found that science fiction writers tend to be a little more accurate than so-called futurists.
Screamers by Phillip K Dick is the short story you are thinking of.
Thank you. I spent 5 to 10 minutes online trying to find the reference. My memory is not what it was :-).
All good. I only remember it because it was made into a movie starring the guy who played “Robocop”, can’t remember his name, my memory is not what it was 🙂