Interim relief is an emergency provision in the employment tribunal regime under which a judge can order a sacked worker reinstated, or at least paid, while their case is heard.
Until recently Britain’s tribunals received about 20 applications a year. They now get about 20 a month in each of 12 regional offices. The judges reckon they know the culprit.
A piece in The Economist last month calls it “agentic flooding” – citizens using AI to submit objections, appeals, and claims against bureaucracies built for the age of the post and the telephone.
The employment numbers are just one example. Single claims rose 39 per cent in the year to March 2026 to 50,000, while the unresolved single-claim caseload rose 55 per cent to 64,000. In some parts of the country, a claim lodged today may not be finally heard until 2030 or beyond.
The lawyers quoted discount the alternatives – there’s no downturn to explain it, and judge numbers and sitting days are above pre-pandemic levels – and land on a technology that locks on to the most obscure provision available “like a heat-seeking missile” and then serves it up to anyone who types a grievance into a box.
Tribunals were built in the 1960s as a people’s court. There’s no fee to bring a case, the losing side rarely pays the other’s costs, there’s no penalty for turning down a reasonable settlement, and judges have held that a badly presented claim still deserves a full hearing if a real grievance might be buried inside it.
One barrister’s conclusion is that there’s no financial mechanism to make anyone think hard about the merits of their case, implying that there ought to be. Tribunals were supposed to make it easier to access justice. And now, when via AI, folk are finally taking the state up on the offer, the system is starting to collapse. It’s like rain on your wedding day, don’t you think.
And just as the technology arrives, the Employment Rights Act adds around 25 new grounds for suing an employer, cuts the qualifying period for unfair dismissal from two years to six months from January 2027, and scraps the £123,543 cap on compensation.
The government’s own forecast is a 17 per cent rise in disputes. Nobody in employment law believes it’ll be that low.
A people’s court
Anyone in or around university complaints procedures and/or the Office of the Independent Adjudicator (OIA) will be familiar with the list. No fee, no costs risk, no penalty for rejecting an offer, and a process designed on purpose to be usable without a lawyer.
Then in England, add in a regulator that closed a consultation in July on a new condition of registration whose stated purpose includes making it easier for students to complain, with decisions due this autumn and fairness requirements proposed to take effect the day they’re published.
The OIA’s own numbers already start to look like the shape of that tribunal curve, just without the acceleration. It received 4,234 complaints in 2025, the first year over 4,000 and a 17 per cent rise on 2024, on the back of nine consecutive years of increases. October 2025 was the busiest month in its history at 516. Where 2008 saw 42.5 complaints per 100,000 students, 2025 saw 165.8.
Twenty per cent of complainants got a favourable outcome. And there’s a capacity signal hiding in the eligibility data – the share of complaints the OIA couldn’t look at rose from 16 per cent to 20 per cent, and 70 per cent of those were students who came too early, most of whom had started their provider’s internal process and given up waiting for a decision. Helen Megarry calls the cases she sees “the tip of the iceberg”.
Scotland, where student complaints go to the Scottish Public Services Ombudsman rather than the OIA, is further along the curve. The SPSO is now warning that most complaints will wait nine months just to be allocated to a reviewer, on the back of a 31 per cent rise in 2025–26 and numbers that have more than doubled since 2019–20 – and unless a student is vulnerable, the matter is ongoing, or there’s a public interest, they wait. In the case of most international students, until their visa runs out.
Beneath the adjudicators there’s no published data at all, because internal complaint volumes sit in annual reports to senates and academic boards that almost nobody outside the institution can read.
What there is, this year, is anecdote – complaints folk and SU advice services telling me that volumes are up, that submissions are longer, and that they arrive citing more grounds than they used to.
It’s anecdote rather than evidence for now, but it’s also exactly how the tribunal story started, with judges noticing something odd well before the problem was formally quantified.
Measuring the reservoir
A flood needs somewhere for the water to come from, and OfS has spent the last 18 months mapping it. Two pieces of research it commissioned, read together, describe a student body that experiences a great deal and formally expresses almost none of it.
Public First polled 2,001 students at OfS-regulated providers in England for a report published in June 2025. Seventy per cent identified at least one disruption to their studies. Twenty-three per cent said they’d received lower-quality teaching than they expected. Forty-two per cent were dissatisfied with how their institution responded to cancelled or postponed teaching, and 45 per cent with its response to industrial action. Just under a quarter said the promises their university made had been fully met.
Savanta’s survey of 1,256 students, published by OfS in January, found 52 per cent had noticed cost-cutting, 83 per cent saw a gap between what they believed they’d been promised and what they got, 54 per cent said the gap was significant, and 56 per cent had never heard of a student protection plan.
Set that against what students do about any of it. A small minority had lodged a formal complaint. Among those who had raised something, informal channels ran at twice the rate of formal ones – 23 per cent through a rep or a conversation, 11 per cent through an actual procedure.
When Public First asked what the biggest barrier was, 36 per cent doubted it would make a difference and 35 per cent lacked the time or energy. A quarter of those who hadn’t complained weren’t confident they’d know how.
Forty per cent said they couldn’t describe their rights at all, and 52 per cent felt poorly informed specifically about their right to compensation. Sixty per cent didn’t know how to complain outside their institution, 8 per cent had heard of the OIA, and of those who complained internally and were dissatisfied with the result, 6 per cent went on to it.
Then the focus groups supplied two barriers the poll didn’t mention:
I think people are scared of getting struck off their course.
I think [going to the OIA] would have to be a pretty serious thing to do, and I think that because it’s external to the university, I’d feel a little bit like a snitch.
To be fair, students couldn’t always reliably separate what they’d been promised from what they’d absorbed as received wisdom about university. A good chunk of the reservoir is disappointment rather than breach, and wouldn’t survive contact with a complaints procedure even if every barrier vanished tomorrow.
But whether we call it under-expression or under-complaining, what’s coming will release both kinds of grievance into systems already creaking under volumes.
Chatbot meets passivist
A couple of years ago now I wrote for SUs about consumer research on passivists and nudniks – the finding that most people who suffer detriment do nothing, and that the small minority who complain regardless of cost are what actually disciplines a market.
The literature gives five reasons people stay passive – the time it takes, an aversion to conflict, a lack of the confidence and cultural capital to assert yourself, fear that the organisation you depend on will treat you worse afterwards, and, before all of those, not knowing the right exists. The OfS research put numbers against them.
A large language model has the capacity to dissolve the first, the third, and the fifth completely – if nothing else, it removes most of the drafting effort that made the first so heavy. Students will soon be able to type “my supervisor stopped replying in February” and within a minute will have the applicable regulation, the procedure, the deadline, a structured letter, and a paragraph on remedy.
The Economist reporters tried it with a fictional grievance and got a claims form, a model legal argument, and tips for cross-examination, plus the unprompted suggestion that the claimant had experienced distress.
What the LLM can’t fix is futility and fear. So rather than a flood, the immediate effect will be a much larger pool of students who will reach the gate marked “is it worth it and is it safe”, and stop there.
What converts that into a flood is evidence that complaining works, and it doesn’t need to come from students. The paper from the Centre for the Governance of AI and the Hertie School that gave the phenomenon its name catalogues 84 surges across 11 jurisdictions and points to the incentive that matters most – an established business model in which intermediaries handle the interaction with a public body in return for a cut of whatever entitlement results.
Claims management firms don’t hold futility beliefs or fear retribution. They hold a percentage. And higher education has the two features that attract them – a large population with similar facts and plausible compensation figures. A cohort whose course closed, or a year group hit by the same industrial action, is a list of leads.
Which makes the OIA’s group complaint figures worth a second look. In 2025 it received 17 of them, from just over 200 students in total, and none was large enough to trigger its large-group process. The collective route through which a thousand identical grievances become one case is sitting almost entirely unused, in a sector where many of the facts of grievance caused by cuts are collective by design.
The shoe shop, revisited
I’ve told the story before of the gap year I spent on the tills in a shoe shop in Wolverhampton, and of the customer who’d turn up weeks later with a worn pair of brogues and a refund demand. What made me fetch the manager was two things together – she seemed to know her rights, and she could cause a scene.
The first worked as a signal precisely because it was expensive. Knowing your consumer rights in 1994 took effort, effort was costly, and so a customer who’d made it was more likely to have a real grievance.
When the appearance of knowing your rights costs nothing, it carries no information, and the rational response of the person on the other side of the counter is to stop treating it as evidence of anything.
That’s what the tribunal judges are now doing. Recent rulings describe submissions running to hundreds of pages, with scattergun grounds, hallucinated statutes, and Magna Carta thrown in for luck. One claimant arrived with 67 grievances over 282 pages, told the judge he intended to rely on a tenth of them, and couldn’t say which.
Another left the judge with the strong feeling that she was pursuing a claim she didn’t understand and couldn’t justify when asked. The models tend to flatter. They tell every claimant the case is strong, because chatbots are biased against telling you that you’re wrong.
Arbel and Shapira’s original theory of the nudnik has a section on what sellers do in response to serial complainers – disarm them. The nudnik’s power rested on being rare and credible. Mass-produced nudnikry is neither, and ends up producing triage rather than introspection. The tragedy of the commons in higher education’s version is that the students with the strongest cases will end up queuing behind the ones with the longest documents.
Nobody asks if it’s lawful
An SU officer told me this week that her university had cut library opening hours. Her instinct was a petition. If those hours were advertised (they were) and they were material to a student’s choice (for many students, the late slots are often the only usable ones) then withdrawing them mid-contract is a consumer protection issue rather than a suggestion for improvement.
Student representation is for things that are basically fine and could be better. You don’t take a broken washing machine to a consumer rep and hope it comes up at the next meeting.
And the sequencing could be actively harmful, because complaints procedures run to windows of a few months – so a term spent petitioning politely and waiting for a committee to say no would deliver a student to the complaints dept out of time – which is the “you should have raised it at the time” rejection that appeals then hand back.
The interesting question is why cuts like that get made without anyone asking whether they are lawful, and I suspect the answer is arithmetic. Take a change affecting a thousand students. OfS polling says roughly half could describe their rights at all. Around a third have ever complained about anything. Six per cent of the dissatisfied go on to the OIA.
Multiply those together and a provider is looking at a handful of external findings, most of which will be worth hundreds rather than thousands – of 4,234 complaints last year, 165 students received £5,000 or more. Against a six-figure saving on library staffing, the expected cost rounds to zero.
Every other option other than the cut has either a large number or a named enforcer attached – insolvency, a missed recruitment target, a redundancy consultation with a real tribunal at the end of it, a league table position. Consumer law has none of that.
Three ways out
Of course, a provider facing rising complaints has three options – it can make complaining harder, it can keep the promises it made, or it can make smaller promises.
The first is what the flooding literature says governments do, and the most uncomfortable finding is that the fastest way to stop a flood turns away the people who need help most. In higher education it would look like page limits, tighter eligibility gates, shorter windows, mandatory structured forms, additional stages, more aggressive routing of complaints into appeals and appeals into complaints, and growth in what gets deemed vexatious.
Each might be defensible on their own as a proportionate response to volume. Each lands hardest on the 36 per cent who doubted it would make a difference and the 35 per cent who lacked the energy, while the student with a model in their pocket sails through – because another forty pages costs them nothing. The passivists of 2025 become the passivists of 2028, the queue rationed by stamina.
The second costs money nobody has. The third is the easiest, which is why it’s the likeliest and the least discussed.
The mechanics are familiar – indicative rather than guaranteed contact hours, illustrative modules, delivery mode and location subject to change, wide variation clauses, reasonable steps rather than outcomes, facilities subject to availability.
The Public First data explains why it works, because if the felt promise is mostly received wisdom rather than contract, hardening the contract downwards costs almost nothing in recruitment terms while removing most of what a well-drafted complaint could hang on.
Consumer law is about whether you kept the promise you made and whether the terms were fair. It has less impact on the provider who promises little and delivers it.
What that does is shift the load off consumer rights and onto the quality assurance system – the B conditions, and the regulation of them. A system which is simply not working, and one that is fatally compromised by OfS’ decision to combine student regulation with fiduciary regulation by binning off the QAA.
There’s a fourth option, which nobody chooses and which is what happens by default while a provider decides between the first and the third. Make the same promises, don’t keep them, and let the queue do the rationing. The OIA’s too-early figure suggests that one’s already underway.
What C6 does
OfS’s proposal for a new condition on treating students fairly uses the research above as its evidence base, and it does so in terms that reverse the sector’s usual reading of a low complaints number. Paragraph 67 cites the 40 per cent who couldn’t describe their rights and draws a blunt conclusion:
Such students are unlikely to be able to exercise their rights by complaining.
The next para cites the confidence and the time-and-energy findings and then goes further than either study did:
We consider that the inherent power imbalance between a provider and a student means that some students may find it difficult to make a complaint due a student’s dependency on a provider for progression, fear of adverse consequences or limited confidence in the decision-maker’s independence from the subject matter of the complaint.
That’s the regulator treating the passivist factors as something a provider will be responsible for dismantling. The bit that follows, C6.4a.iii, would require all reasonable steps to give students timely, clear, and effective advice on complaints and redress, and specifies that this means free, independent, and confidential advice throughout the process.
Elsewhere, providers would be expected to make clear public commitments that a student who complains won’t be penalised “academically or otherwise (for example, through withdrawing financial support or visa sponsorship)” – that’s aimed at the one barrier no LLM can touch, and at the international postgraduate for whom it bites hard. And the publication requirements would put complaints documents and an explanation of the OIA’s role on a single page, against a baseline of 8 per cent awareness.
Two things aren’t there. There’s no expectation on learning from complaints – nothing requiring a provider to analyse what comes in, and nothing about the pile of institutional critique that arrives inside academic appeals and disappears when the appeal fails on academic judgment.
The OIA says that in more than two-thirds of the academic appeal cases it saw last year, students raised other concerns about teaching, supervision, feedback, or pastoral support alongside their personal circumstances, and that material vanishes. Ignoring or dismissing concerns, failing to investigate properly, and delaying redress appear only in an illustrative list of things that wouldn’t satisfy the good-faith principle.
And there’s nothing at all that constrains the friction response – not a word on page limits, eligibility gates, shortened windows, or vexatious designations. The condition proposes to push hard on the demand side but leaves providers free to build the dam.
The poor could win too
The Economist’s leader concedes that at first sight, the flood looks like a brilliant result for fed-up citizens, and that the poor could come to exercise their rights as successfully as the sharp-elbowed middle classes already do. I’ve argued on here before that rights are rationed to those with the stomach or the resources to fight for them. What’s arrived is a technology that will help to un-ration them.
The mature student who could never have afforded a solicitor, the international master’s student who didn’t know the OIA existed, the commuter whose inter-campus bus vanished in a budget round – for the first time each of them can produce a competent, referenced letter in an afternoon. That’s a proper access-to-justice gain, and a sector that talks about students as partners ought to want it.
But the tribunal evidence says the gain gets squandered unless three things sit alongside – knowing where the line between a grumble and a breach sits, complaining in the right lane and in time, and knowing when a hundred identical complaints ought to be one.
The Economist’s endgame is that every worker ends up with a top-flight lawyer in their pocket and a deluge of slop gives way to a wave of winning claims. The higher education version turns on whether SUs are the thing that converts the one into the other.
The regulator has rightly identified that the gap between what students experience and what they complain about is a gap it wants to be closed. AI will accelerate the process whether it likes it or not.
What’s undecided is which direction it closes from – from below, with students learning where the line sits and students’ unions helping them hold it, or from above, with providers redrawing the line so that less falls on the wrong side of it.
One of those involves bulletproofing, legal advice and liability limiting. The other takes a proper partnership with students and their SUs that respects students’ rights, resolves their legitimate complaints and delivers improvements. Which the sector reaches for first will tell students all they need to know about whose side their university thinks they’re on.
