.
Artificial intelligence made the biggest steps forward in regard to perception, but also good progress on decision-making. It appears to be feasible to create a small autonomous robot that can accomplish a mission set by 2030.
The combination of autonomous robots with miniaturization may yield autonomous robots the size of small birds with useful performance.
Small autonomous drones may be very practical for quantity production. Let's assume that such drones become very effective militarily. The result may be that military power in the 2030's may depend first and foremost on the ability to acquire by production or import millions if not tens of millions of autonomous war drones.
This may spell the end for main battle tanks and other line of sight weapons platforms save for high-flying aircraft: A couple thousand or ten thousands of anti-tank guided missiles were incorrectly predicted to make the concept of main battle tanks obsolete, but millions of tank-penetrating drones would almost certainly make line large and expensive of sight combat platforms obsolete.
What would such autonomous war drones look like?
A basic concept would be a kamikaze killer drone. In concept it would be a kind of missile, albeit with loitering and / or waiting ability, possibly also the ability to recharge itself from the environment. It would detect and identify a target, decide to engage, and likely be consumed by the hit. A shaped charge warhead would be used to penetrate vehicles at known weak spots, and a much smaller EFP-style shaped charge could be used to kill individual troops without the need to close in with it the last few metres.
We will no doubt devise and test, if not deploy, numerous countermeasures ranging from netting to drone-hunting drones. So we can expect drones to replay the growing specialisation of military aircraft. Initially, there were observer aircraft, but the urge to kill was great enough to develop ground attack aircraft almost immediately. This created the need for fighter aircraft. Later we added passive electronic reconnaissance aircraft, active electronic warfare (jamming) aircraft (even different types for different wavelengths), refuelling aircraft, transport aircraft, target practice aircraft, decoy aircraft (typically unmanned) and so on.
We might see such a wide range of autonomous drones as well, and in addition to the task specialization we might also have terrain-specialised types. Drones for urban settings, drones for open fields and drones for woodland could benefit from terrain specialisation for low detectability and endurance. Particularly well-equipped military bureaucracies might also include war drones that could submerge in water and thus hide in lakes, ponds and slow-moving rivers - and exploit the underwater stealth for an approach movement. Other drones may be tasked as stay-behind sabotage drones, and hide under leaves, foliage or even in soft soil.
Operational art may become about layering detection and defence against drones, organising supply & stocks, and deploying large swarms of mission-tailored composition to search, probe, push, occupy and finally meticulously clear out defined areas.
Concentrations of drones would be engaged by classic artillery, but other than that the classic trio of infantry, tanks/cavalry and artillery may be replaced by drones.
The first area to become dominated by drones might be mountainous areas, for the classic trio is the least capable there. Artillery has difficulties to hit reverse slopes, infantry move extremely slowly and tanks cannot negotiate much of the terrain.
Meanwhile, drones could handle such terrain quite easily, though flying drones may require enlarged rotors or wings to compensate for the low density atmosphere at altitude.
The last area to become dominated by drones might be urban terrain, as buildings offer many opportunities to restrict the movement of small drones and an army would want to interact with the civilian population rather than just kill.
Autonomous drones would be a novelty, and both the conservatism of the armed bureaucracies and the political desire to not open Pandora's Box too much might delay the build-up of stocks of such hardware. Rapid technological progress might also render stocks obsolete in a few years, possibly reducing their value to that of decoys. That would be another disincentive to a stocks build-up.
The armed bureaucracies would have much disdain for the drone business, as it wouldn't be soldierly, wouldn't follow what the officers were taught when they were till young, armies would need to think like air forces to grasp the matter, the need for most career officers and their skill sets would vanish and so on.
There are thus many reasons to expect a rather slow build-up, even slow R&D. That is, except in nations whose devious leaders seize on such a land warfare revolution as a means to topple the balance of power and succeed with some war(s) of aggression.
Personally, I see little reason to fear robot wars more than human troops wars. We certainly proved that the latter can turn genocidal and easily break societies, nations, if not cultures. To send out drones to kill is no more anonymous than the movement of division markers on Field Marshall Haig's table or the firing of a 155 mm HE shell by a howitzer crew. War drones won't be able to reproduce themselves, so their threat to mankind would be limited by their numbers even if they turn against their human masters.
There's still the Star Trek-ish scenario of fully automated (or simulated) war becoming all-too easy, but don't we have this problem with wars abroad already anyway?
War might finally cease to be much about morale, and become all about readiness and money.
related:
/2008/01/screamers.html
/2009/06/will-5th-be-last-manned-fighter.html/2013/06/killer-robots-and-loal-munitions.html
/2014/02/drones-in-theory-part-1-introduction.html
/2014/02/drones-in-theory-part-2-aerial-drones.html
/2017/08/very-low-level-air-defence-against.html
/2018/01/an-open-letter-to-fellow-pacifists.html
/2018/07/link-drop-and-comments-july-2018.html
S O
.