The US Pentagon has announced its new Replicator Program, a plan to counter China’s numerical superiority in troops, weapons and especially drones with thousands of inexpensive autonomous US drones, due to arrive within 18 to 24 months of Deputy Secretary Kathleen Hick’s announcement on August 28.
This program has many implications, but there is one key change we’ll see running through not only weapons but many technologies going forward.
A lot of basic tech can be readily upgraded by AI, and will be.
And in the case of “autonomous weapons,” that means a lot of drones operating with little to no human supervision.
Yes, there are a lot of safety implications in such changes, which America and other advanced democracies are concerned about, but most civilians aren’t flying fighter jets, driving tanks or sailing warships, so some missions can readily incorporate safeguards.
To use another AI option, automatically detecting and shooting down a hypersonic vehicle flying directly towards your fleet in the middle of the ocean is fairly straightforward - there are almost no airborne vehicles capable of reaching such speeds other than missiles, especially in the atmosphere, and we don’t have to worry about whether an incoming rocket is authorized and filed a flight plan.
When we get down to the level of ground troops, fighting ground troops, urban warfare and anywhere military forces and civilians may be in close proximity or mixed together, things get much more complicated.
We’ll discuss this in depth as we continue this series, but something to remember about remote control, autonomous AIs and the combination of the two is we’ve had the underlying technology for some time. What’s really happening is it’s getting cheaper, better and much more widely distributed.
So the war between Russia and Ukraine has been featuring vast numbers of drones for both fighting and reconnaissance, but wholly or mostly autonomous weapons still seem to be rare-to-non-existent on the battlefield.
One of the drivers of the AI race, though, is the stated goal of China and Russia to dominate this field and to weaponize it, and Beijing has been working for years towards the “intelligentization” of their military - the full integration of AI into all aspects of warfare.
This is grim, to be sure, which is why there’s more of an effort to head it off by limiting China and Russia’s access to Western technology - such as advanced semiconductor chips - than simply to deter it with more-advanced systems, though obviously the US, NATO and other allies are working hard in that domain as well.
Many systems to counter these autonomous weapons will come into play, such as jamming, spoofing sensors, concealment, disruption of the logistics bringing them to the conflict zone and finally fighting them.
As their range and sophistication improves, however, they will become harder and harder to stop, likely leading to a technology gap many nations will be working hard either to expand or to close, depending on which side of they’re standing.
We’ve discussed some of the limitations of AI weapons previously, and will certainly do so again. Hacking, disrupting networks, striking invisibly or faster than they can process and constantly changing what they’re dealing with, even in the space of seconds - all of these tools can blunt the effectiveness of the first AIs we’ll see in combat.
But again, preempting their acquisition and use is probably the best initial tactic, and working with close friends and allies to insure no one feels comfortable attacking you with them is likely the second
.