Too close for missiles, switching to guns
Too close for missiles, switching to guns
The future of dogfighting
The latest instalment of my technology newsletter for the Wavell Room - this time on AI in air power. Hope you enjoy, and do sign up if so!
Issue #9 | 12 September 2023 | Edited by Kenneth Payne
As AI masters ‘drone racing’ and the Pentagon bets heavily on massed swarming, what future for fighter pilots?
Image Credit. TOPGUN Maverick.
Too close for missiles, switching to guns
Do you feel the need for … dogfighting? Maverick did, of course. But why worry about it today in an era of autonomous aircraft and high-tech missiles that can hit targets far beyond visual range? On Twitter/X recently there was an interesting exchange on exactly this issue. The Economist’s defence editor Shashank Joshi replied to a tweet about AI dogfighters:
So why dogfight? – surely it’s an anachronism. I thought about that Twitter exchange this week, when I saw a clutch of AI stories. In the first, I saw that an AI had competed well in drone racing against elite human pilots. No surprise, perhaps – AI has already defeated experienced pilots in simulated aerial combat, and a real-world test using converted F-16s is in the pipeline. Next, news that the DoD is going all out to field very large drone swarms in the near future - Emma had more on that in last week's newsletter. Having deep pockets, however, the Americans are hedging their bets. This week also saw news of USAF plans to acquire more than a thousand Valkyries – the somewhat larger UAV from Kratos which looks like it will find employment as a ‘loyal wingman’ for piloted F-35s and NGADs.
That hedge illustrates a high degree of uncertainty about concepts and weapon design for aerial AI. The unique selling point of AI is its decision-making ability, relative to humans (even aircrew ). AI brings speed and accuracy, which as aficionados will know are important ingredients in Boyd’s OODA-loop decision-making model. Since you can clone them more readily than you can acquire pilots with the right stuff, algorithms should enable mass too. Hence the swarm, and the idea of saturating targets. Make thousands of cheap, disposable platforms, and who cares if you lose a few hundred in combat. The F-35 in Shashank’s tweet can fire off all 8 of its AIM-120 Sidekick missiles. The Valkyries alongside can swing into action too. No matter – my drone swarm has many hundreds more aircraft than they carry missiles, and is closing fast.
Of course, there’s more to it than that. There are many, many trade-offs to consider in weapon design. What’s the range, airspeed and endurance of my small swarming drones? How robust are they in stormy weather? How much firepower can they deploy, given their size? Perhaps they’ll be mid-sized, a bit like these Northrop Grumman V-Bats that Shield AI demonstrated this week, part of its research on swarming. Perhaps they’ll be more like a cloud of manoeuvrable missiles– something like these tiny Perdix drones from MIT, publicly demonstrated ages ago now when filmed emerging from pods beneath a crewed F-18. With no crew to preserve, all manner of possibilities emerge. Should you favour speed, manoeurvrability, or stealth? Payload or deployability? Should you spend money on exotic materials and aerodynamics, or is it the hive-mind code that’s the real value added?
Two things strike me about this debate.
One – the lesson from earlier technologies is that no matter how radical and exciting your new gizmo, it’s the combination of different platforms that brings real combat effectiveness. If you forget about combined arms warfare, you’re liable to be embarrassed by adversaries that don’t. So, in Ukraine, it’s abundantly clear that ubiquitous tactical drones are having a big impact on combat – but also that they’re doing so in coordination with established combat arms – especially artillery.
Two – turkeys don’t vote for Christmas. Which is to say that culture matters in shaping how technologies are developed. That’s broad cultures, at the societal level, where there’s a degree of unease about the ethical implications of AI, including in the application of lethal violence. But also, narrower cultures, whether of militaries, air forces, or aircrew. There’s a cockpit on the mock-up Tempest, which may say something about the present technical limitations of AI, but also – perhaps – something about the proclivities of those designing the platform.
Meanwhile, I saw this week that the Prince of Wales headed out to sea this week. Finally, there’ll be F-35s flying off the second British carrier. But the ship is also sailing with a variety of UAVs, as the Royal Navy, like everyone else, looks ahead to a future with many more uncrewed aircraft. Will the robots buzz the tower? Anyone’s guess.
Some more air power developments below.
IN BRIEF.
Cardboard drone ftw
For a couple of thousand dollars a copy, these are certainly cost effective. Ukraine has apparently used flatpack drones supplied by Australia to destroy a clutch of Russian aircraft and some air defence installations. Apparently there are plenty more where they came from, but let’s hope it doesn’t rain.
Return of the Bayraktar
You knew that it would be back. Signature weapon, alongside the Javelin, of the early exchanges in the Russia-Ukraine war, its return to the battlefield is an indicator that Russian air defences aren’t what they were.
Read more.
Challenger 2 destroyed.
And one from the land domain. It looks like the inevitable has happened – the destruction in combat of the first UK supplied Challenger 2 tank in Ukrainian service.
Do you have tech or science material you want us to cover? Reach out through our contact form here.