UK MoD braves the weather to train maritime AI capabilities Britain has conducted a maritime trial to collect data for developing AI technology for the Ministry of Defence. While large chunks of the country eyed the weather forecast nervously, an exercise led by the Defence Science and Technology Laboratory (Dstl) sent personnel out onto Hampshire beaches in what the MoD described as " …
Hardly what I would call large when you only use 130 the MoD may describe it as "one of the largest maritime artificial intelligence data capture trials.", but that just shows how small the UK's armed force is, or how "little Ai studies" have been carries out.
That... seems a little premature. The exercise wasn't driven by data, it was to collect data. And since that's only just happened, it's questionable whether AI has uncovered any insights, yet. If indeed it ever will, since this exercise is described as data collection for training AI, rather than AI analysis of said data.
So it's another minister talking out of their arse, again. No clue what they're yammering on about, they just think is sounds compelling. Business as usual, sadly.
"That... seems a little premature. The exercise wasn't driven by data, it was to collect data."
I suspect this trial and data collection was influenced by previous trials, rather than being the first ever one of its kind. So probably yes, a data-driven data collection activity.
Mm, dont worry. In the case of AI lets hope the government keep their fingers out of the pie and just provide the money and some kind of ethical framework. When I was working on Skynet Remote Monitoring and Control (RMC) software, (yes, Skynet does exist but not in the form Arnie would recognize) we simply had the remit to make it work better than if a human was in control.There were n ethical remits, it was just fancy tracking software.
it's questionable whether AI has uncovered any insights, yet. If indeed it ever will
The British military have been using AI for years. As in big computers wading through loads of data to try and produce some meaning out of it. Given that real AI doesn't actually exist.
For example, back in the dim and distant days of the 70s/80s military intelligence were getting overwhelmed by the amounts of ground radar data they got from radar reconnaissance planes. You could see all these military vehicles wandering all over the shop - but what were they doing? So apparently they used the fast forward and rewind feature on video recorders to see where convoys of trucks were going (and coming from) and used that to work out the enemy's logistics. Where is that big old pile of Soviet ammo so we can make it go kaboom? This then became a subject for early computerisation. And I believe they've been using machine learning to try and improve this sort of stuff for decades.
I believe they've also been using machine learning in trying to discrimate radar targets over bumpy land, or water (which is both bumpy and moves) is quite difficult. Plus trying to pick up stealth aircraft. Also in determining submarines from general sea sounds. But this tends to be an area where nobody is telling you what they're doing. Plus I guess you need to distinguish between the use of increased computer power to do clever stuff and actual machine learning.
They haven't got an "AI". Various people have got different kinds of "A?". Perhaps an ASG - artificial statistical guesstimator - or ABG artificial bullshit generator. On which subject I was speaking to a salesman the other day who told me that his product was in the top right corner of Gartner Magic Quadrant. It was only on video chat. I wonder if he was artificially generated? I've never seen a human mention these without immediately mocking them before - I assumed it was a sort of mythical sales technique only ever seen in parody.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
