One to watch for in next week's Budget and associated forecasts is the suggested impact of AI on employment rates in the UK.
The FT has reported this morning that:
Law firm Clifford Chance is cutting about 10 per cent of its business services staff in London, pointing to increased use of artificial intelligence as one reason for the job losses.
They added:
Roughly 550 employees, in areas including finance, HR and IT, were told last month about plans to make some 50 jobs redundant and to bring in role changes for up to 35 others.
And to contextualise this, they added:
The elite firm, where partners earned an average of £2.1mn in the past financial year, told staff that greater use of AI and reduced demand for some business services meant that it needed to cut jobs.
Let's be clear: we always knew law was an area where AI would have a significant impact, and I suspect many firms will be doing what Clifford Chance is. Other sectors will not be hit as badly. But to ignore this issue would be reckless on Rachel Reeves' part. But will she really address it? I am not betting on it.
AI: How to use it as a campaigner
We have produced a guide on how to use AI as a campaigner, for which we're already getting good feedback. It's available as a free download here. Take a look. We think you'll find it useful.
Comments
When commenting, please take note of this blog's comment policy, which is available here. Contravening this policy will result in comments being deleted before or after initial publication at the editor's sole discretion and without explanation being required or offered.
Thanks for reading this post.
You can share this post on social media of your choice by clicking these icons:
There are links to this blog's glossary in the above post that explain technical terms used in it. Follow them for more explanations.
You can subscribe to this blog's daily email here.
And if you would like to support this blog you can, here:
Buy me a coffee!
Two thoughts spring to mind
I remember when I started work, almost all records were paper and there were huge numbers of staff employed dealing with it.
Thats all gone now and all documents are held electronically BUT its a process that took over twenty years
Secondly I am reminded of Admiral Fisher, I was looking at two of Drachinifels documentaries about him last night. He went to sea on a ship basically the same as HMS Victory but went on to build HMS Dreadnought and served on the committee developing ASDIC.
He was a technophile and recognised the potential of the Submarine and the Diesel Engine BUT it took a long time before the full potential of both these technologies were to be fully developed.
Might I suggest that we are in a similar situation to the one he was in just before WW1 there are new technologies out there, in this case AI BUT there is a way to go yet before they are working at their full potential (Air blast fuel injection anybody?) so I might be a bit more cautious.
Thanks
The consequences of AI will be felt in the next Budget rather than this one. That will be when Reeves or her successor have to find the money to pay Universal Credit to everyone who lost their job to a LLM.
The same AI that’s decimated the work staff at HMRC. Unless you are an accountant, it’s practically impossible to call them anymore and actually talk to a human. A merry-go-round of ineffectual messaging that doesn’t provide answers.
I suspect the problems at HMRS predate AI. When i had them as I client in the early 2000’s I recall that they had around 90,000 staff. Now they are more like 60,000 with local offices that dealt especially with SME’s closed. As Richard has pointed out before HMRC more than pay their way in tax recovered.
On merger in 2000 there were 100,000 staff.
Now about 65,000
I am waiting for the AI produced demands for the Winter Fuel payment claw back via HMRC with interest. No point in employing AI when you have a Chancellor who can’t decide what she wants, when it’s going to start and which year it will affect. I don’t expect the prompt to be made public.
I am sure that the legal profession is standing on the edge of an unemployment precipice. AI is already having an impact in litigation (an area that many had supposed would be less affected because of the need for someone to turn up to court and argue a case) where pleadings and research is increasingly undertaken by AI (sometimes with humorous but unintended results). However it is in the transactional areas of the law – commercial contracts, property transactions, finance agreements etc – where, despite the mythology surrounding them, most of the tasks can be automated. In reality they have been for decades with the use of standard precedents. The intellectual input involved was reduced to just selecting the right precedent clause for the right agreement but now that input can be done by AI. The news about Clifford Chance”s announcement is a warning. Clifford Chance (Clifford Turner & Coward Chance) was the first mega merger in the 1980s – where it leads the rest follow. Who would want to be a law student now?
Most law students I have met of late are now primary school teachers, working like crazy, but it’s better than law.
No doubt their vital M&A work will continue. I see they are currently advising on the acquisition of a UK milkshake company called Shaken Udder! But I hope their pro bono work isn’t affected.
I am assuming that all these lost posts are not forward facing with direct contact with the “public”
A better use of AI would have been how to make savings with the minimum job losses and a suggested strucuture to support it. It may well have suggested a reduction in the high wages of some which could then be used to reduce costs without any redundancies.
AI is only as good as the prompting it is given and a simplistic prompt will give a simplistic answer ” cust staff costs”
I can remember starting work as an office junior in 1972.
There were two men employed to literally read the same stuff into a dictaphone on a daily basis, and two women employed to type up said same stuff.
This was in a quite small family business in Shoreditch, London.
I bet they would have been pleased with AI!
“I bet they would have been pleased with AI! ”
People generally aren’t that pleased when they lose their job. Funny, that.
As part of this, Clifford Chance is also moving roles to or creating roles in their lower-cost offices in Newcastle, Warsaw and India. I suppose creating jobs in another part of the UK, rather than London, is a good thing, though not much consolation to those losing their jobs in the capital.
AI is often used as an excuse because firms like Clifford Chance like to look “cutting edge.” There is always another reason too, the real reason, that is slightly less glamorous or even embarrassing.
What is actually happening is that Clifford Chance wants to reduce its costs by offshoring some of the work. In turn, this shifts some of the HR etc offshore too. People don’t tend to like offshoring though, so they claim it is AI. I also notice a drop off in business, which makes sense because Reeves triggered a recession with her last budget.
LLMs are still incredibly limited in what they can do and a firm like CC wouldn’t dare rely on the output because of the volume of hallucinations. I suspect the lawyers are having the same experience as I have. It can save a bit of word smithing that means a bit less overtime, but that is about it.
LLMs are very reliable for legal work, I would suggest.
This type of law is about constructing documents from pre-defined choices, not asking questions. AI could be very good for that, and populating the documents with data. This is just an advance in database design.
I would counter argue that it’s not really doing any more for form filling in my area than VBA could already do. To some extent the input has decreased but the review has increased. Granted, the applications are now more stable but I still think the long term effect of AI will be a reduction in skills premium for certain employees rather than outright unemployment.
However, the loss of skills premium should in theory make a lot of things much cheaper… we will see.
If AI can be “by and large good enough” – by which I mean, not necessarily perfect, but not more likely to be wrong than a real person – then it may have the positive impact of opening up access to the law (and other professional advice and services) to those for whim it would otherwise be beyond their means. But that may mean one person’s AI talking to another person’s AI with neither of the clients having any clue whether they are getting things wrong or right. Would you trust it, if you didn’t know better?
I give AI the facts I want it to use.
I tell it what I want.
I tell it why I want it.
I tell it how I want it.
And then it usually does pretyty much what in want and can save me editing time.
I do not trust it much beyond that.
According to its website, Clifford Chance has 719 partners, and over 2,500 other lawyers. Its profits per equity partner are over £2m. The loss of 50 support staff (out of 500) is not really moving the needle – taken together, they may be paid about the same as one or two partners – but it might be an indication that restructuring of jobs and tasks and roles is on the way.
It has ever been thus. The fax room is no more, the post room gets much less traffic, no more filing clerks, very few firms have a coffee or tea trolley, the typing pool is almost gone, hardly anyone gives or types up dictation, and so on. AI is reviewing documents – bundles of leases or land registry records, or due diligence materials, or disclosure in litigation – and producing first drafts of reports. I read yesterday of a firm using an AI tool to screen job applicants with immersive recruitment exercises. Everyone is looking for use cases.
I thought AI might displace us quite rapidly as welfare rights advisers … potentially. However, I have found that in some instances, we have to “defend” our advice to people who have consulted ChatGPT or Google AI and got the wrong answer to their question on benefits rules and how they apply to their circumstances. My concern right now, is whether my job will now require that I demonstrate that as an adviser (with a fair amount of experience), I know better than AI when I spot the errors …. 🙂 It’s already happening. May be just a matter of time and AI will catch up! I doubt it will acquire the compassionate touch to defending those facing hardship though!
Much to agree with
I’ve already had meetings where I have had to demonstrate (using books!) why I am right and ChatGPT is wrong. Clients are already saying “you are not telling me something I could not find out for myself” – to which my answer is: good luck with relying on that. I am qualified, I have a regulator, and insurance. You are buying peace of mind through access to my experience. But if you’d rather make your own furniture and clothes and build your own house and do your own accounts and tax returns and draft your own contracts and and defend yourself in court, well, good luck.
AI may catch up but I’m less certain about that than I was – LLMs may be too poisoned with falsehoods already. But perhaps they could learn to have some humility.