OMG. AI hallucination in police intelligence report used to decide on Maccabi's game

[PerkingFaintly]

It's like all the possible disasters rolled into one!

https://www.bbc.co.uk/news/live/c394zlr8e12t
11:54
What did the report say about the West Ham match that didn't exist?
The intelligence report that referred to a non-existent match between Maccabi Tel Aviv and West Ham has not been published in full.
But it was referred to by Lord Mann during a Home Affairs Committee session on 1 December.
He said: "Early on in the intelligence report, it says: 'The most recent match Maccabi played in the UK was against West Ham in the Europa Conference League on 9 November 2023. This was part of the '23-24 European campaign. It marked Maccabi Tel Aviv’s last competitive appearance on UK soil to date.'
"That is in the intelligence report, but that did not happen. West Ham have never played Maccabi Tel Aviv.

[1145]
Chief constable's survival hangs in the balance
For [Chief Constable] Craig Guildford it is hugely embarrassing to have to admit that his officers did use AI - after he had personally told MPs that they did not.
That use of Microsoft Copilot led to information going into the report that referred to a game between West Ham and Maccabi Tel Aviv that had never actually happened.

As a siren voice saying how shite "AI" is for the past 5 years, do I win a prize ?

Slightly different issue, but more breaking news on "AI" just now: Musk has reportedly caved and banned Grok nudifying women and children.

Still awaiting confirmation and clarification.

https://www.theguardian.com/politics/live/2026/jan/14/pmqs-live-digital-id-cards-keir-starmer-kemi-badenoch-labour-conservatives-latest-news-updates

Bloody hell😮

From that article:

UPDATE: Starmer said:
I have been informed this morning that X is acting to ensure full compliance with UK law. If so, that is welcome, but we’re not going to back down. They must act. We will take the necessary measures. We will strengthen existing laws and prepare for legislation if it needs to go further, and Ofcom will continue its independent investigation.
FURTHER UPDATE: Darlington was referring to this tweet about Grok changing its policy. We have not seen independent confirmation of it yet.

And this is the Tweet referenced:

https://x.com/PolitlcsUK/status/2011284195583410192
@PolitlcsUK
🚨 NEW: X has now banned Grok from generating sexualised images of women and children after the UK made it illegal to create “non-consensual intimate” images

However, it is still responding to requests to put men in bikinis or sexual positions
3:48 am · 14 Jan 2026

I'm just hoping the legal professionals aren't dependent on AI these days. Two friends buying houses. First took seven months, second home so no chain second four months, again no chain, just soo slow, one query at a time.

The Cabinet Office published a report on their cross-government trial of Copilot. https://www.gov.uk/government/publications/microsoft-365-copilot-experiment-cross-government-findings-report/microsoft-365-copilot-experiment-cross-government-findings-report-html

Even in this likely polished up to look as good as possible official report, in short it's useful to save time on routine boring admin tasks but not great for anything complicated. "Limitations were observed when dealing with complex, nuanced, or data-heavy aspects of work."

We did a basic test of Copilot in some training and asked it to generate a list of three real place names where the letters in the name are alphabetical. Then ask it to evaluate that list. It gave three places, then said one wasn't alphabetical, one was fictional, etc so actually the whole thing was incorrect sorry.

AI is a language model - people don't understand what they're using. It's not a fact checker.
It's only as good as it's sources, so if it's looking at junk, nonsense and bias, that's what you'll get. At the cost of many litres of water to keep enormous data centres ticking over.
High profile cases like this are necessary to show up the flaws and limitations.

High profile cases like this are necessary to show up the flaws and limitations.

Which won't matter a hill o'beans. Too many people have spunked too much money into tulips "AI" for it to be reined in now.

It could be the tobacco of the 21st century. Something "everyone" is into that turns out to be deadly.

🎖️

Yes - I don’t use it normally, but a few times I have put technical documents through an AI for summaries just to see what it would come up with. Sadly, the summaries sound very plausible — until you read the original document, and find that the summaries don’t actually bear much relationship to the original. So if you rely on AI summaries, you most probably just never know how you’re getting things wrong. Same for other AI I’ve seen - it all sounds so good, unless you actually know a lot about the subject, and then you can tell how sketchy and sometimes downright wrong it is.

And yet, on other threads about AI we will be constantly told it’s the future of the workplace and we must all get with using it, it’s a better therapist than a person, etc. When reading those threads I just think everyone’s gone mad and is affected by some kind of hallucination themselves. It’s just not very good. I don’t know why anyone would actually rely on it at work? You’d just always be getting stuff wrong and never even know it?

This absolutely! Everyone knows AI can hallucinate, so it always needs to be dealt with sceptically & really challenged critically if decisions are being based on it.

Everyone ?

My favourite is when my clients come to me and say "oh I don't think we have tax to pay because ChatGPT told me..."

I can't, because policy, but if I was running my own practice I'd tell them if they're prepared to take that risk then I'll charge them double to fix the resulting mess

Well more folks are catching on

Never ever underestimate the ability of someone whose job depends on black being white telling you that you need to listen rather than look.

Too many people are plums deep in "AI" to stop it now. Your best bet is to develop compensating controls where you can.

First the judge in the Sandie Peggie case using it, now a senior policeman. Ffs.

Its particularly concerning when you consider the internet is overrun by bots.😱

There[s already been a case where a kid was trying to get high safely and died of an overdose due to AI hallucinated guidelines

That would be shocking had I not watched a President of the United States suggest injecting bleach to cure Covid.

Read a story on sky news about monkeys on the loose somewhere in the US. Attempts to track and capture them were being hampered by people using AI to show the monkeys were "spotted in their garden" etc.

AI is just the next nightmare in the downward spiral :/

Are you sure they don't mean "ICE agents" ?

It reminded me of that judgement too. People will be more careful if it costs them their job.

@1apenny2apenny Who is higher up than a chief Constable in terms of operational policing? They call the shots and if politicians told him what to do, they should be removed and he should have called them out.

The trouble is that governments and regulators everywhere are either deliberately or because they're congenitally slow about a lightyear behind the people who have put money into AI.

Deliberately = countries (not just the USA btw) who want their stock markets to show bubble-like growth fueled by AI to counter inflation and economic stagnation, and/or who are in thrall to tech bros.

Congenitally slow = anyone who's worked in any public sector role anywhere would know that "nimble" isn't typically applicable, except in high stakes military and diplomatic capacities.

This is a really, really thorny thing to address. You'd ideally want a council of ethicists, scientists, lawyers, regulators, finance bros, tech bros, educators, healthcare professionals etc etc etc to convene globally to think all this stuff through and come to a conclusion. In reality, not only has the horse already bolted, it's way over the horizon heading in a direction that literally nobody can guess.