The UK’s leading combat air companies and the Ministry of Defence have collaborated with scientists and engineers in the field of machine learning, artificial intelligence, data science, and computing to develop software for new fighter jet Tempest.

This project is part of the UK’s future combat air system (FCAS) initiative.

Tempest is designed to be a supersonic stealth fighter, featuring advanced integrated sensing and protection technologies. It relies on extensive software systems, both on the aircraft and in ground-based systems.

“These capabilities will be delivered, in part, by millions of lines of code on the aircraft, with many more lines of code also present in ground-based systems. This means the software on Tempest needs to be more robust and resilient than that on its potential adversaries. The collaboration provided valuable insights into software requirements, design, delivery, operation, speed of upgrades and maintenance for both the fighter jet and the training systems pilots and maintainers will use to operate and support the aircraft. 

Outsmart Insight, a deep tech intelligence company, and Oxford Creativity, a group delivering a systematic approach to innovation and creative problem solving, conducted targeted research with scientists, engineers and academia. The research addressed the most challenging problems facing software development over the expected multi-decade life of the programme: flexible ways of managing computing resources; the role of trusted artificial intelligence; software re-use; and increasing software dependability.”

Air Commodore Lowe, FCAS Programme Director for the MOD, was quoted as saying:

“Software is key for Tempest because the future operational environment demands adaptability, including frequent software updates. But software is also a big delivery risk. Recent history shows the dangers that arise when software is done badly and the advantages of doing software well. The advantages are so significant that, in terms of operational capability, the people delivering the software are as important as the people maintaining the aircraft or the pilots flying them.”

Tempest is targeted to be in service by 2035.

Avatar photo
Lisa has a degree in Media & Communication from Glasgow Caledonian University and works with industry news, sifting through press releases in addition to moderating website comments.
Subscribe
Notify of
guest

82 Comments
oldest
newest
Inline Feedbacks
View all comments
Coll
Coll
3 days ago

The MOD must be creaming themselves with this new buzzword bonanza; machine learning, artificial intelligence, data science, and computing to develop software.

Last edited 3 days ago by Coll
SailorBoy
SailorBoy
3 days ago
Reply to  Coll

Well, Tempest is supposed to use loads of AI in construction and service. Probably how BAE got the contract to build it

Coll
Coll
3 days ago
Reply to  SailorBoy

I’m not having a pop at the construction process. I’m having a little giggle at the MOD must be creaming themselves over new buzzwords.

SailorBoy
SailorBoy
3 days ago
Reply to  Coll

Obviously, I was pointing out that BAE probably used the fact to help gain the money for the contract

Coll
Coll
3 days ago
Reply to  SailorBoy

Ok. No worries.

SailorBoy
SailorBoy
3 days ago
Reply to  Coll

Sorry, that was a bit abrupt, the obviously was saying that your statement was correct I didn’t mean to be disparaging

Coll
Coll
3 days ago
Reply to  SailorBoy

Nar, you are good. It’s more about how MOD PR can be embarrassing.

Spyinthesky
Spyinthesky
3 days ago
Reply to  Coll

Not clear why it’s embarrassing to use actual words that describe actual technology. ‘Data Science’ is probably the closest to a buzzword whereas Machine learning is no more one than human learning. Mod are certainly prone to using buzzwords which to my mind tend to be obscure in their meaning but these ones as I say above are predominantly descriptive of more recently developing technologies,

Mark B
Mark B
3 days ago
Reply to  Coll

Interestingly there is little that is new about those buzzwords or companies claiming new wonderful methodologies which will fix everything.

Coll
Coll
3 days ago
Reply to  Mark B

I was having a bit of fun MOD PR. I am of defence companies and their buzzword corporate speak. I’m sure if you banned them from saying these term they would be lost. It’s just like using words like efficient, cost-effective, reversibility, effective, scalability, diverse, survivability, combatant, commonality, etc.

Last edited 3 days ago by Coll
Spyinthesky
Spyinthesky
3 days ago
Reply to  Coll

So what term would you use in place of ‘machine learning’, or ‘artificial Intelligence’? They aren’t made up, they are the actual terms used for those technologies and seem clear to me.

DaveyB
DaveyB
3 days ago
Reply to  Coll

More words for meetings bingo sheet

Spyinthesky
Spyinthesky
3 days ago
Reply to  Coll

Just descriptive I would say.

AlexS
AlexS
2 days ago
Reply to  Coll

In one paragraph it justified 10 years delay and a multibillion pounds increase in spending.

Jim
Jim
3 days ago

This will be great for the Chinese ambassador to the UK that said last week the UK is not a competitor to China on AI because it’s so small. Love to see an AI powered UK aircraft shoot down whatever piece of shit Russian knock off the supreme leader and his horde can create.

Coll
Coll
3 days ago
Reply to  Jim

Check out ‘China Fact Chasers’ for some funny AI China videos.

Mac
Mac
3 days ago
Reply to  Jim

The more the Chinese go out of their way to slag a country off, the more worried they are about that country behind closed doors.

They spend alot of time slagging off Britain.

Spyinthesky
Spyinthesky
3 days ago
Reply to  Mac

Indeed they always like to put down those they know are at least in some regards a technological threat. Fact is we punch well above our weight even if we can’t compete on numbers with Countries far bigger than us. There’s a reason so much of the modern World came from UK based minds.

Mr Bell
Mr Bell
2 days ago
Reply to  Mac

Agree- the Chinese know that British inventiveness, innovation and science will surpass any technology that China can steal through industrial espionage or reverse engineering. The key aspect of technology seems to be computational power- so quantum computing and <6nm microprocessor blocks. As long as the west has an advantage and in fact superiority in this field the Chinese know their weapon system technology will be inferior. The troubling aspect around China is their ability to mass produce equipment/ ships and weapon systems- especially as it is Western companies that have supported Chinese industrial expansion. China is estimated to have 4x… Read more »

Jim
Jim
2 days ago
Reply to  Mac

Russians too, juts look at how that worked out for them.

Spyinthesky
Spyinthesky
3 days ago
Reply to  Jim

Considering DeepMind is one of, and arguably the most advanced AI Company on the planet the Chinese Ambassador as usual is talking through his ass. We don’t have the mass the US or China has in the sector that’s true, but like our F1 engineering, at the top end it is as good as anyone’s and I would wager more advanced than Chinas. As an example DeepMind’s AI has just raised the bar existencially on the number of stable crystalline structures identified, hundreds of thousands indeed which are already starting to be created and tested in labs for their potential… Read more »

Barry Larking
Barry Larking
3 days ago
Reply to  Jim

The U.K.’s entire bloated university system depends on Chinese students. None of these students is studying sociology or sports facility management I wager. But you point is well made – and colourful!

MrSatyre
MrSatyre
3 days ago

AI: “Yeah, Tempest is okay, but would be better with canards.”

Human: “Nooooooooooooooooo!”

Toby J
Toby J
3 days ago
Reply to  MrSatyre

No, no, no
Tempest looks just fine the way it is, and won’t need superb manoeuvrability to win battles. Canards would just get in the way

Spyinthesky
Spyinthesky
3 days ago
Reply to  Toby J

It’s laughable how the YF 22 beat out the YF 23 when its only advantage was extra manoeuvrability for a dog fight thanks to its weighty nozzle design, because those pesky Russians already had it. In reality before it even got into service that factor was already becoming of minor relevance as over the horizon missiles became de riguer. Fact is if an aircraft ever has to use vectored thrust these days plan A and B will already have failed and plan C no guarantee of survival. The YF 23 would have been a far better bet using its superior… Read more »

M knowledge
M knowledge
3 days ago
Reply to  Spyinthesky

Very true

DaveyB
DaveyB
3 days ago
Reply to  Spyinthesky

One of the big factors that decided the YF23’s fate, was that it wasn’t in the same fly able condition as the YF22. Lockheed at that stage were notorious for delivering products late. Northrop bent over backwards to get a flyable demo aircraft ready before the deadline. To put things in perspective the YF23 was only just behind the YF22 in angular pitch rate, ie how fast it can move its nose up or down. This was due to its V-tail. Where each eleven had near enough the same surface area as a F18’s main wing. One advantage that the… Read more »

Robert Blay
Robert Blay
2 days ago
Reply to  Spyinthesky

If that competition took place today. The YF23 would win because of the reasons you highlighted. Both were born in an age when who could pull 9g the longest was seen as still the most important aspect of modern air superiority. How times have changed.

Spyinthesky
Spyinthesky
3 days ago
Reply to  MrSatyre

AI is developing aircraft that don’t need vertical stabilisers so I doubt they will be contemplating canards at least when the emphasis put to them is on stealth.

MrSatyre
MrSatyre
3 days ago
Reply to  MrSatyre

Apparently the fine folks here don’t grasp the joke here that AI is not infallible. Tsk.

Jon
Jon
3 days ago

ChatGPT has insisted on the necessity of real furry dice in the cockpit rather than a virtual set. The Furry Dice Requirements Committee meet next Thursday in closed session in the hope that they can avoid strike action from the Minimalist Cockpit Group. A minimalist spokesperson went on record saying that AI is too stupid to understand just how dangerous furry dice could be in a 9g turn, and how a complaint will be lodged with the Tempest Feng-Shui Subcommittee. Last night, despite back channel messaging between the two parties, no compromise deal had been reached and experts would not… Read more »

Marked
Marked
3 days ago
Reply to  Jon

The furry dice will be delayed and it’ll be fitted for but not with for the first decade of service.

DeeBee
DeeBee
3 days ago
Reply to  Marked

Yes, you just know it, that’s if ( and it’s a big if) it ever gets built at all

Coll
Coll
3 days ago
Reply to  Marked

It’s delayed on the basis if they want traditional or DnD die.

Last edited 3 days ago by Coll
Jack
Jack
3 days ago
Reply to  Jon

As long as it has Playboy Bunny mud flaps behind the landing gear, I’ll be happy.

DeeBee
DeeBee
3 days ago

Tempest is Targeted to be in service in 2035 🤔🤔!! Any more good jokes 😂😂

Geoff Roach
Geoff Roach
3 days ago
Reply to  DeeBee

For once I think we might actually make it.

DeeBee
DeeBee
3 days ago
Reply to  Geoff Roach

For what it’s worth I hope we do, other countries with nowhere near our military aviation history/ expertise are doing so, so In theory there’s no reason why we can’t make a success of Tempest, but if there’s one thing this country can balls up on an epic scale its usually a project like this.🤞

Geoff Roach
Geoff Roach
3 days ago
Reply to  DeeBee

Sadly we have had more than our fair share of cock us but I’m hoping with Italy and particularly Japan onside we’ll do it this time. 👍

Geoff Roach
Geoff Roach
3 days ago
Reply to  Geoff Roach

cock us? I’ll try that again. Cock ups.

RoboJ1M
RoboJ1M
3 days ago
Reply to  Geoff Roach

And no French this time! 😂

Mr Bell
Mr Bell
2 days ago
Reply to  RoboJ1M

Agree stay clear of the French- keep them at arms length, ditto Saudi. Get the aircraft into service and then see the foreign military sales take off. Result: big win for British industry and the balance of trade.

Jon
Jon
3 days ago
Reply to  DeeBee

We hope to have the tech demonstrator around 2027. I can believe that. In fact I’d hope to see it rolled out as early as Christmas 2026, but without all the bells and tinsel. Will we see a prototype multi-function-radio-frequency-doodad in place by then, or “just” ECRS Mk 2? Full adaptive engines? I’d hope so. To have the first fighter in service by 2035, eight or nine years later is also believable. However IOC I’m having trouble swallowing. I think we can do the technology, but I’m not convinced we can commit to getting a full squadron of fighters built,… Read more »

Louis
Louis
3 days ago
Reply to  Jon

‘a true 6th Gen system’ What is that? As is commonly said, this generation saying was made up by LM to sell 5th gen fighters. 4th gen and 5th gen have become easy to distinguish mainly via stealth characteristics. What will be the difference between 5th and 6th gen? The main 6th gen projects are NGAD, F/A-XX, SCAF and Tempest. NGAD is a project of just 50-100, 200 max long range air superiority fighter. It doesn’t need to have any air to ground capability because the USAF will also have 100-200 B21s, and 1800 F35As. F/A-XX will replace the circa… Read more »

Jon
Jon
3 days ago
Reply to  Louis

Gen 6 isn’t the Tempest fighter alone. It’s the interplay of the fighter and other systems around it. It’s the ability to not only select your missile load-out for the mission but your supporting drone platforms too, knowing that your core fighter will be able to control them all with AI assistance and without cognitively overburdening the pilot. Squadrons will be considerably less homogenous. I think you are right and Tempest will be complementing 5th Gen (when available), also a lighter AEW platform that will have lost most of its current control function. Just as you wouldn’t expect to send… Read more »

DaveyB
DaveyB
1 day ago
Reply to  Jon

Bearing in mind, there will also have to be carrier variant of SCAF! Which will be a French only affair, are the other three Countries expected to cough up funding for its development as well?

Spyinthesky
Spyinthesky
3 days ago
Reply to  Louis

Good question it’s nebulous, the main differentiators being the sensors, awareness, Ai based cooperative capabilities and flight control, electronic complexities (including power generation to feed it), superior stealth and adaptive engines but there are any manner of ways to include/exclude such factors or massage their capabilities when included. In many ways they are just more advanced sets of what some Gen-5 aircraft have, esp when comparing to the F-35 which has set a high bar and will be further improved. As for 4th/5th Gen and the stealth characterisation as differentiator I just read that (and I can’t confirm its veracity)… Read more »

Geoff Roach
Geoff Roach
3 days ago

STOP PRESS…AI to replace MOD procurement office. 😉

Coll
Coll
3 days ago
Reply to  Geoff Roach

I was going to make a joke that the AI software would have a MOD option that considered MOD procurement decision making.

Last edited 3 days ago by Coll
Geoff Roach
Geoff Roach
3 days ago
Reply to  Coll

We’ll share it Coll 😀

Coll
Coll
3 days ago
Reply to  Geoff Roach

lol

Stc
Stc
3 days ago
Reply to  Geoff Roach

If it did Mr Roach the Chinese would not be saying the UK was too small then. If it replaced humans on a like for like basis even the yanks will be embarrassed by the size of ours !

Jack
Jack
3 days ago
Reply to  Geoff Roach

I.N.T.E.L.L.I.G.E.N.C.E from Team America could outperform UK MOD in procurement.

Nath
Nath
3 days ago

Hope they get rid or the tails and take the magna approach.
High Mach, low observability.

But I quite like the idea of changeable covers. Like the old Nokias. One for deep strike on for dogfighting.

M knowledge
M knowledge
3 days ago
Reply to  Nath

I want a private company to buy the tsr2 and modernize it lol

DaveyB
DaveyB
3 days ago
Reply to  M knowledge

Seriously, you’d be better off using the Vulcan, as a foundation for modernization. As it would be more easily adapted to rebuilding from a basis of a very good low observability. Along with being a pretty decent stand-off cruise missile carrier. For instance as part of a fun exercise to make a stealthy Vulcan from an existing airframe, implementing 10 fairly easy modifications: 1. Place louvres in the intakes, to shield the engines. Intakes will probably need a redesign if money is available to include S-ducting. Along with including radar absorbent tiles in the intake ducting. The intake’s vertical boundary… Read more »

Armchair Admiral
Armchair Admiral
3 days ago
Reply to  DaveyB

YES THE vulcan. Lovely semi stealth plane. Great you tube videos about Vulcans Nuking the USA in excersises.Only because the Banshee drone
Looks a bit Vulcan_y…it has allegedly a very small radar cross section (it’s not very big for a start), but mostly because it’s fibreglass…how difficult would it be to reshape the panels to make it truly stealthy? I mean, once you have the new mould and a bit of low_viz paint on it….would it not then become very useful.
Sorry, just a random thought.. AA

DaveyB
DaveyB
2 days ago

Hi AA, not a stupid thought at all! The Qinetiq Banshee must fly with a radar reflector, which not only reflects radar but also amplifies it. This makes sure that a ship’s radar for instance can detect it as it pops over the horizon. The repeater helps the Banshee mimic an anti-ship missile’s radar cross section (RCS). The Banshee also uses flares and high powered IR lights to enhance its IR signature. When these enhancements are switched off. It has been known that lookouts spot the Banshee before the ship’s radar does. You are quite right that it is mostly… Read more »

M knowledge
M knowledge
23 hours ago
Reply to  DaveyB

As much as I think that would work I think the TSR2 would be the ultimate platform. The fact that it was already amazing it wouldn’t need many tweaks.. literally upgraded with scramjet technology and modern day EW kits. A potential Mach 4 or 5 monster.

Spyinthesky
Spyinthesky
3 days ago
Reply to  Nath

Be interesting to see if NGAD eliminates them as many claim. Plenty of experimentation going on but like vectored thrust there are potential compromises to be had that probably the quality of coding will probably be the tipping point.

M knowledge
M knowledge
3 days ago

Here in the UK we are doing very well. I just fear Labour getting in and cancelling our hard work.

PeterS
PeterS
3 days ago

” Millions of lines of code” is a recipe for precisely the huge cost overruns and interminable delays that have bedevilled the F35 programme and continue to limit the effectiveness of the aircraft.
Tempest won’t be funded as generously as the F35 so the risks of failure and cancellation are much greater. This programme needs to have clear milestones for taxpayers money to be provided.

Spyinthesky
Spyinthesky
3 days ago
Reply to  PeterS

Yes an argument not to be dismissed. Will just say that the F-35 is greatly held back by its 90s/Noughties base software origins. Much has been learned since then, above all modularity one example being separating weaponry control from the flight control (to simply the comparison admittedly) so that updating the former doesn’t mean complete re-testing at great length of the latter. How well this new era of programming works in practice however will be the prime indicator of whether 6th Gen development is more reliable than 5th Gen. and whether it enables core software and OS often ancient in… Read more »

magenta
magenta
1 day ago
Reply to  PeterS

and yet … Air Commodore Lowe, FCAS Programme Director for the MOD, was quoted as saying: “Software is key for Tempest because the future operational environment demands adaptability, including frequent software updates. But software is also a big delivery risk. Recent history shows the dangers that arise when software is done badly and the advantages of doing software well. The advantages are so significant that, in terms of operational capability, the people delivering the software are as important as the people maintaining the aircraft or the pilots flying them.” (my emphasis) Windows 11 has about 50 million lines of source… Read more »

Ralph Hardie
Ralph Hardie
3 days ago

Will the ship fill a crucial role in 2035?
The Typhoon is totally useless for the Russian invasion of Ukraine. The Gripen currently seems the better choice. But artillery shells by the millions would be best.

AlexS
AlexS
2 days ago

Tempest is useless. Its airfield will be destroyed.

Just make a missile with long range, lets say 4000km . If it needs to be in a vector for an extra range, an Airbus 320 can certainly fire it from 4000km distance. The missile performance can be improved every 5 years or even less.
Tempest will be stuck in the airframe for 50 years, so it is a serious limitation. Instead if needed you just make a different missile.

Mr Bell
Mr Bell
2 days ago
Reply to  AlexS

Sacrilege, how would the RAF ever be able to justify its pampered fly boy image without an airfield to fly those glorious flying machines from?
In reality a lot of what you state is true- airfields are all air forces Achilles heels, dispersion and STOVL capabilities are key to keep jets deployable and able to fly regardless of airfields being hit/ taken out.
Of course a missile might not get through if the airfield had some rudimentary GBAD, which the RAF regiment is clearly lacking.

SailorBoy
SailorBoy
2 days ago
Reply to  Mr Bell

Tempest should have been STOVl, it would look so cool

SailorBoy
SailorBoy
2 days ago
Reply to  AlexS

The whole point of Tempest is that it is modular and has room for advancements over its lifespan. I’m not sure what point you’re trying to make; we should scrap actual combat aircraft and fit airliners with cruise missiles? Really? Because that’s fine, as long as your (enormous to get to 4000km) missiles can’t be shot down, are immune to jamming, the enemy doesn’t know your coming to put up air patrols, your missile is accurate enough at 4000km to hit a target the size of a hangar. That’s before: if your missile is supersonic, the chance of missing is… Read more »

AlexS
AlexS
2 days ago
Reply to  SailorBoy

Modular is just pep talk. You are always stuck with airframe. Tell me how many years and expense has been necessary to just fit Meteor in F-35? These super complex aircrafts are preposterous when a missile and a good recon can do the work. It would have to be a family of missiles – sort of CAMM for anti-land with various engines for diverse range and speed, loitering and diverse warheads or decoys- that can be fired from land, submarine, aircraft/drone. A missile always will have an advantage of not being shot down over a much bigger aircraft. Why a… Read more »

SailorBoy
SailorBoy
2 days ago
Reply to  AlexS

I would refer you to the 1957 Defence Paper, which proposed the inverse of what you say (missiles for defence, rather than offence), widely regarded as the worst procurement decision until the 2010 SDRP (or whatever) that wrote off TSR2 among others
The main point against your suggestion is that a combat aircraft can be used for purposes other than flattening buildings. Most of what Typhoons do over Syria involves gathering intelligence and target identification for ground troops. An airstrike is a last resort

SailorBoy
SailorBoy
2 days ago
Reply to  SailorBoy

Though, as you mention it, a modular Aeralis style family of missile engines, fins, seekers and warheads would be an interesting concept. You could have super-long range AAMs, land attack, ASMs… With common parts, launch methods and factories. THAT would be a good idea. Relying on it for defence, not so much

monkey spanker
monkey spanker
1 day ago
Reply to  SailorBoy

Please don’t compare anything useful to that bongadoodle aerialis.
I will believe that when it flies. Modular aircraft🤦🏼‍♂️

SailorBoy
SailorBoy
1 day ago
Reply to  monkey spanker

I know, I know, but same concept? I wish Aeralis had worked, would have made a real difference to training style

AlexS
AlexS
1 day ago
Reply to  SailorBoy

You just need intelligence gathering drones, satellites, even said missile with a proper engine module and head module can have a loitering function to do recon for the attacking ones.

AlexS
AlexS
1 day ago
Reply to  SailorBoy

You are quite right the 1957 Defence Paper was a disastrous one do to its precocious emphasis into missiles. Precocious is the operative word.

But technology moves on and what was no possible to achieve reliably can be possible after enough developments.

SailorBoy
SailorBoy
1 day ago
Reply to  AlexS

Ok, let’s say you’ve convinced me. The UK is going to replace its fixed air wing with AEW- and Ground-launched long-range (Ballistic?) land attack, Anti-ship and AA missiles that can be transported on lorries to the point of use. Presumably Apache would also be replaced with LPS, Surface Brimstone or similar. Where do you start? Get BAE to fill out a concept launcher for a range of launchers and VLS for ships? Get the MBDA cogs whizzing with a fat £5B for the range? Come to think of it, a land-based FC/ASW wouldn’t be a bad idea, but the logistics… Read more »

AlexS
AlexS
1 hour ago
Reply to  SailorBoy

Apaches only make sense if they will be much cheaper, otherwise it will be drones. At Operational level i would start with 10 and 30km missile in Spike class to destroy all enemy tanks before they even approach our infantry lines. If they have to use a Javelin it is already failure. Then guided 60, 100km, 200km etc in various range increments. The advantages of ground based is that is more difficult to detect and can be even fired unmanned and from underground and cheaper, disadvantage is that if you are out of range it takes more time to be… Read more »

SailorBoy
SailorBoy
9 seconds ago
Reply to  AlexS

I agree with you on the small scale, Spike/drones. Also re OPV/Corvette and SSK for sea lanes. Also the ability of transports to launch long range missiles. However, there is still something to be said for the ability of fighter jets to invade protected airspace and hit fleeting targets of opportunity, something missiles are simply incapable of

zzhxy
zzhxy
2 days ago

i think GCAP should use adaptive engines, FRENCH and NGAD already decided to use adaptive engines

SailorBoy
SailorBoy
1 day ago
Reply to  zzhxy

I think it already does. BAE website would say more but can’t say for certain off the top of my head.

DaveyB
DaveyB
1 day ago
Reply to  zzhxy

That is what Rolls Royce have alluded to. As well as incorporating the in-shaft generator and pump system. The test engine that was shown some time ago, judging on appearances and measuring it against a person standing underneath. It looks quite a bit wider in diameter than the Typhoon’s EJ200. Not to mention that Tempest when it is actually revealed, will be quite a bit bigger than Typhoon, due to the need to include large weapons bays and more internal fuel storage. Therefore to maintain at least a performance parity with Typhoon, this new engine will need to develop considerably… Read more »