Please or to access all these features

Work

Chat with other users about all things related to working life on our Work forum.

Thread gallery
7
blueshoes · 21/10/2026 22:04

In my job (risk analysis), we still need to keep the human-in-the-loop. I think AI requires more human judgment to churn through the torrential output, not less. GenAI is brilliant at finding and putting data together at speed, a task done by admin staff. Hence in my team, once we have incorporated AI processes into the workstreams, we will need half the admin staff but to be replaced by more analysts which can (be trained to) interact intelligently with the AI output.

I don't see a future for clerical, admin, processing staff. Similar to the role of secretaries becoming practically obsolete in my sector, once agentic AI is introduced in my firm, I suspect it will wipe those remaining roles out.

The only thing slowing this change is that the senior people don't have time to fully engage with learning how AI can streamline tasks. Entire work processes can be re-imagined but it is not in the interests of more junior staff to be transparent and engaged with this transformation because it means them losing their jobs. Hence we are still in this delicate stalemate but the tipping point will come.

One day, it will come for my job too. But I will be retired by then. Just fear for my dcs.

RaggedRocks · 21/10/2026 22:05

I ask AI for racing tips.

PerkingFaintly · 21/10/2026 22:11

Well I've just been reading about "LLM grooming", so yes, pretty worried.

"LLM grooming" is the publication of disinformation on websites optimised for web-crawlers and chatbots, rather than human readers, to manipulate chatbots to use the disinformation in their summaries. It's successful because humans so rarely click through and check the chatbot's sources.

We already have the situation where many people believe any old shit they see on the internet.

LLM grooming makes this an order of magnitude worse. The chatbots act as a cut-out between the reader and a blatantly dodgy source, effectively laundering the disinformation.

Russia is ahead of the game in this... as far as we know. But it's a game anyone can play, so...

Claude and Mistral Echo Kremlin-Aligned Claims in Over 70% of Tested Answers
https://united24media.com/investigations/claude-and-mistral-echo-kremlin-aligned-claims-in-over-70-of-tested-answers-23655
[Global Disinformation Index (GDI)] linked the findings to the rise of zero-click searches, which it put at around 60%, where users accept an AI summary without visiting the original source. “Whoever controls the AI answer summary will control the narrative,” GDI concluded.

Claude and Mistral Echo Kremlin-Aligned Claims in Over 70% of Tested Answers

The GDI study finds large language models facilitated Russia disinformation by drawing from Kremlin-aligned sources in most tested data voids.

https://united24media.com/investigations/claude-and-mistral-echo-kremlin-aligned-claims-in-over-70-of-tested-answers-23655

blueshoes · 21/10/2026 22:11

I am not convinced AI will cause mass unemployment. For office work, it will allow people to focus on the more interesting aspects of the job. However, not everybody wants this. Some people prefer to 'shovel snow' but that is not going to be a realistic option in an AI world. People who previously would have become admin clerks will have to find something else to do.

RaggedRocks · 21/10/2026 22:13

I am not saying AI picks me winners, but it saves me the bother of reading all the tipsters' analysis. I never bet more than a pound, but it entertains me.

Jk987 · 21/10/2026 22:16

@MermaidMummy06 your DH didn’t provide the right requirements for the holiday then. That’s not an issue with the AI.

Aussiesgettingsmashed · 21/10/2026 22:34

RaggedRocks · 21/10/2026 22:05

I ask AI for racing tips.

Are you a millionaire?

RaininSummer · 21/10/2026 22:34

It's going to cause a lot more issues than mass unemployment I think.

Nktes · 21/10/2026 22:35

Jk987 · 21/10/2026 22:16

@MermaidMummy06 your DH didn’t provide the right requirements for the holiday then. That’s not an issue with the AI.

I've found it unreliable for these things too. I asked it to plan a journey for me and it thought there was a direct train, but it actually meant an underground and then a train.

blueshoes · 21/10/2026 22:36

PerkingFaintly · 21/10/2026 22:11

Well I've just been reading about "LLM grooming", so yes, pretty worried.

"LLM grooming" is the publication of disinformation on websites optimised for web-crawlers and chatbots, rather than human readers, to manipulate chatbots to use the disinformation in their summaries. It's successful because humans so rarely click through and check the chatbot's sources.

We already have the situation where many people believe any old shit they see on the internet.

LLM grooming makes this an order of magnitude worse. The chatbots act as a cut-out between the reader and a blatantly dodgy source, effectively laundering the disinformation.

Russia is ahead of the game in this... as far as we know. But it's a game anyone can play, so...

Claude and Mistral Echo Kremlin-Aligned Claims in Over 70% of Tested Answers
https://united24media.com/investigations/claude-and-mistral-echo-kremlin-aligned-claims-in-over-70-of-tested-answers-23655
[Global Disinformation Index (GDI)] linked the findings to the rise of zero-click searches, which it put at around 60%, where users accept an AI summary without visiting the original source. “Whoever controls the AI answer summary will control the narrative,” GDI concluded.

@PerkingFaintly I have wondered about this risk and see that this is now true with LLM grooming.

If it is possible to manipulate Google's algorithms in the form of search engine optimisation (SEO) to rank higher in Google's search results, then someone is likely to be able to figure out how to manipulate the public LLMs to feature in their answers. I have asked ChatGPT to recommend an investment and at the back of my mind, wondered how they arrived at those core investments.

Now bad actors like Russia are in on the act using it to manipulate public opinion.

This is dangerous as I feel there is still quite a high level of public trust in LLMs like ChatGPT and Claude. If we cannot trust LLMs, then what they produce is just slop but many people will lack the critical thinking to challenge the answers. I include myself in that category as I often ask ChatGPT for things that I don't know enough about, not for things I have enough core knowledge to know ChatGPT is feeding me bilge.

PerkingFaintly · 21/10/2026 22:45

I often ask ChatGPT for things that I don't know enough about, not for things I have enough core knowledge to know ChatGPT is feeding me bilge.

Cogently put. That's one of the dangers in a nutshell, isn't it?

RaggedRocks · 21/10/2026 22:46

Aussiesgettingsmashed · 21/10/2026 22:34

Are you a millionaire?

Not yet

Gottagetta · 21/10/2026 22:50

Soso3 · 21/10/2026 20:39

My workplace has already stopped hiring for some roles due to introduction of AI. I work in finance and we are finding that a lot of work can be automated and the people who leave aren't being replaced. The whole team is worried and reckon they could end up with just 2 employees out of 7 originally.
We are also worried about the job becoming monotonous, the AI is taking on more and more work and those left will just end up doing two of three tasks on repeat.

Edited

My husband is an FD and he recently told me that pretty soon AI will be able to do his job for him.
“But they’ll still need people at the top to decide what they want the AI to do,” I said. He said the people “at the top” will be the people who are the most skilled at using AI. He’s an experienced finance professional not an AI expert.

ilovepixie · 21/10/2026 22:56

Namechangedatheist · 21/10/2026 20:44

I've worked in IT for 40 odd years. I suspect that AI is a complete game changer. It is already replacing large numbers of jobs and I think this will escalate very quickly over the next few years. Massive social change is almost inevitable and this is never easy.
This is without considering the very real risks from a massive increase in cyber attacks by 'bad actors' using AI. I was at an industry cyber security conference last week and one of the pieces of advice was to remove all your household appliances (fridges etc.) from the internet...
Added to the possibility of a 'super-intelligent' AI going rogue I think we are completely stuffed.
I have honestly never felt so depressed about the future at any point in my life.

Why what are our appliances going to do?

BurntBroccoli · 21/10/2026 22:58

HouseHouseHouse7 · 21/10/2026 21:28

Sam Altman, the CEO of OpenAI, said something along the lines of …some jobs will go away and new ones will be created, the shape of jobs will change.

Im hoping he’s right but I don’t know enough about the topic.

Im closer to 60 than 20 so it won’t affect me, but I do worry a bit for my young adult kids and stepkids.

The issue is that it is happening very quickly and the new jobs just haven’t had time to even be invented yet.

When you think about things like typing pools which were pretty much obsolete after 50 years with the advent of the computer, I think AI will change things in about 5 if not sooner.

Supperlite · 21/10/2026 23:02

Terrified. It feels apocalyptic.

PerkingFaintly · 21/10/2026 23:03

ilovepixie · 21/10/2026 22:56

Why what are our appliances going to do?

Well our fridges could all heat up to 50° C overnight and spoil all our food, so that everyone loses their food simultaneously, creating shortages, and perhaps leading to civil unrest.

Or for a real win, maybe some of the fridges might catch fire.

This month's cyber attack on supermarket fridges didn't get fully that bad, but it's a taster of things to come.

Fridges in French supermarkets hit by cyber attack
https://www.bbc.co.uk/news/articles/crq5nzxv9q4yo
Hundreds of French supermarkets have been hit by a cyber attack that caused their fridges and freezers to heat up, spoiling fresh and frozen food estimated to be worth millions of euros.

Hackers gained control of the software that regulates the temperatures of the refrigeration units, forcing some to warm to 50C (120F) overnight, according to one supermarket chain affected.

Intermarché - one of the two chains known to be affected - has more than 2,000 stories across France but only a proportion of those were hacked, reports suggest

Fridges in French supermarkets hit by cyber attack

Hundreds of stores were targeted by the overnight hack, which spoiled fresh and frozen food estimated to be worth millions of euros.

https://www.bbc.co.uk/news/articles/crq5nzxv9q4yo

Waitingfordoggo · 21/10/2026 23:05

ilovepixie · 21/10/2026 22:56

Why what are our appliances going to do?

Well, there was a cyberattack recently on some AI- enabled freezers in a French supermarket. They turned the temperature up or switched them off or something so the shop lost loads of stock. It’s easy to see why hackers target businesses, less clear why they would target private citizens, but if it can be done then it probably will be because so many humans are just shitbags. I suppose hostile nations might do it just to cause general havoc and financial loss on a large scale in their enemy countries.

Waitingfordoggo · 21/10/2026 23:07

Supperlite · 21/10/2026 23:02

Terrified. It feels apocalyptic.

Yep. Can only hope that there are some decent people in the upper echelons of the organisations running the show, cause the rest of us are utterly powerless.

Shatgtp · 21/10/2026 23:08

Well I think we are all cooked.

peckishcats · 21/10/2026 23:09

I spoke to a friend yesterday who told me that his partner has just been let go from an IT job, and will find it very difficult to find another similar role.

Things are definitely concerning.

PerkingFaintly · 21/10/2026 23:10

To be fair, I'm not sure the French attack involved AI, rather than hacking by humans.

It's the fact that appliances are internet-connected that's the problem.

But adding AI features into the mix is another trendy selling point encouraging people to unnecessarily connect shit to the internet, thus making it vulnerable.

Waitingfordoggo · 21/10/2026 23:11

I see what you’re saying @PerkingFaintly , my tech knowledge is terrible so thank you for explaining.

Toutafait · 21/10/2026 23:57

Waitingfordoggo · 21/10/2026 23:07

Yep. Can only hope that there are some decent people in the upper echelons of the organisations running the show, cause the rest of us are utterly powerless.

This article was published last month, and I think it's already out of date:

What to know about 'the singularity,' AI's point of no return

disgracefullyyours · 22/10/2026 00:04

I agree with others about their being a mental health crisis that will happen - take away purpose and income from people and you’ve got a recipe for disaster. Employers in general don’t care that AI may not be as accurate as some of their best employees - it just has to be much cheaper than the cost of employing humans. No benefits to pay it, no NI, no pension, just a one-off fee and maybe a monthly maintenance charge, or a monthly subscription.

Swipe left for the next trending thread