AI is already showing up inside UK employment tribunals, just not in the way many HR teams expect. There is no wave of rulings yet. What exists instead is a single, closely-watched case testing how discrimination law applies to algorithmic decisions, a tribunal system whose own judges say AI is already changing how claims are argued, and two regulators moving well ahead of any case law.
At the Employment Tribunal National User Group meeting in March 2026, the President of the Employment Tribunals told members that judges shared the view that AI was “likely behind an increase in the complexity of claims, applications for reconsideration and applications for interim relief, as well as inflated schedules of loss.” That is not a data point buried in a footnote. It is the tribunal system describing, in its own words, a shift already underway.
This piece looks at what is actually known so far:
- The one UK tribunal case testing AI-driven discrimination claims, and what it did and did not establish
- Why official statistics cannot yet isolate “AI-related” claims as a category, and what that means for how HR should read the numbers
- Where regulators are already active, ahead of any tribunal ruling
- What the available evidence signals for employers right now
Explore: AI Grievances and Employment Tribunals: A Guide for HR Teams
How Many AI-Related Claims Are We Actually Seeing
Tribunal claim volumes are rising sharply across the board:
None of this volume is broken out by cause. There is no official category for “AI-related” tribunal claims, and the judiciary has been candid about why. The President of the Employment Tribunals described the link between AI and rising complexity as “anecdotal,” based on judges currently being able to spot AI-generated content in a document by eye, not on any formal tracking system. That is a genuine limitation on what can be measured, not a sign that nothing is happening.
This matters for how HR teams should read what follows. The absence of an official “AI claims” statistic reflects how early this still is, and why the small number of concrete cases and regulatory actions that do exist carry more weight than their number alone would suggest.
The Claim Categories Emerging So Far
With no official data breaking claims down by cause, the categories below are drawn from the one live UK case testing this area, alongside findings from the two regulators already active on AI in the workplace. These are the areas where a claim is emerging or plausible, not a ranked list of confirmed tribunal volume.
Discrimination and bias. The clearest example is Manjang v Uber Eats UK Ltd, covered in full below. Separately, the ICO’s Recruitment Rewired report, based on engagement with over 30 employers, found recurring gaps in how organisations monitor their automated hiring tools for bias.
Monitoring and surveillance. The ICO has already taken enforcement action here, though not through a tribunal. In 2024, the regulator ordered Serco Leisure to stop using facial recognition and fingerprint scanning to monitor employee attendance, finding staff had no clear way to opt out and no meaningful alternative was offered. It is a data protection enforcement action rather than a discrimination claim, but it shows the same underlying pattern: automated monitoring introduced without adequate consultation or opt-out.
Procedural fairness, specifically the absence of human review. This is the thread connecting almost everything above. Manjang’s case centred on the claim that no meaningful human review took place before his deactivation. The ICO’s Recruitment Rewired findings centre on the same issue from the regulatory side: employers believing they have human oversight in place when, in practice, review is inconsistent or superficial.
The categories overlap more than they sit apart. A single dispute, over facial recognition, an automated hiring decision, a productivity monitoring tool, tends to raise discrimination, data protection, and procedural fairness arguments together, rather than falling neatly into one box.
Notable Rulings and What They Actually Established
Manjang v Uber Eats UK Ltd is the case most often cited when AI and UK employment tribunals come up, and it is worth being precise about what actually happened, because it is frequently overstated.
- 2021: Pa Edrissa Manjang, a Black Uber Eats driver, permanently suspended after repeatedly failing the app’s facial recognition verification checks
- Claim: indirect race discrimination, harassment, and victimisation, arguing the software is less accurate for Black faces and that he was never given a meaningful human review before losing access to work
- May 2022: the tribunal refused Uber’s strike-out application, finding it could not conclude the claim had “little or no reasonable prospect of success”; also allowed Manjang to amend his claim and kept the UK entity as a respondent
- 2024: case reported to have settled, before it reached a full hearing
This is where the case is commonly misdescribed. The tribunal did not rule that Uber’s facial recognition software was discriminatory. It ruled only that the claim was strong enough to proceed past an early strike-out challenge, a procedural threshold, not a decision on the merits. The substantive question, whether the software’s outcomes actually amounted to unlawful discrimination, was never decided by a tribunal.
What the case does establish, even at this preliminary stage, is that a UK tribunal was willing to let an AI-related discrimination claim proceed on the basis that facial recognition software, and the absence of a clear human review process, could plausibly produce a discriminatory outcome. The Equality and Human Rights Commission, which funded the claim, has itself noted that there is currently little case law testing AI and human rights breaches in the UK. Manjang is one of the first cases to raise these questions at all, which is precisely why its procedural outcome, rather than a final ruling, is still significant.
What These Rulings and Regulatory Action Signal for Employers
Taken together, the Manjang case and the regulatory activity around it point to the same handful of concerns, regardless of which body is looking at the problem.
- Meaningful human review is the central issue. Manjang’s claim turned on the absence of a clear human check before his deactivation. The ICO’s Recruitment Rewired report found the same gap from the regulatory side: employers often believe a human is reviewing an automated decision when, in practice, that review is inconsistent, superficial, or applied to some candidates but not others. A human glancing at an AI-generated shortlist is not the same as a human genuinely able to change the outcome.
- Transparency with the people affected. The ICO’s findings and the EHRC’s evidence to Parliament both raise the same point from different angles: people are often not told clearly that an automated system played a role in a decision affecting them, or how to challenge it. Manjang’s case included the argument that he was never given a clear explanation of what process led to his suspension.
- Lack of a genuine alternative or opt-out. In the Serco Leisure enforcement notice, the ICO found staff had no real alternative to biometric monitoring and would likely not have felt able to object given the power imbalance with their employer. That same imbalance sits underneath most AI-related workplace disputes: the person affected rarely has a practical way to say no or to challenge the system before harm occurs.
None of this requires a finished body of case law to act on. Every one of these concerns, human review, transparency, a real opt-out or challenge route, is something HR teams can audit and fix now, well before a tribunal or regulator asks the question directly.
Where This Is Heading
None of the pressure described above is easing off. If anything, it is converging from multiple directions at once:
- The Employment Tribunal National User Group itself expects the trend to continue rather than settle. Judges’ shared view was that AI was already shaping claim complexity and volume, at a time when the tribunal system is already straining under a rising backlog.
- Regulators are moving in parallel, not waiting for tribunal case law to catch up. The ICO’s consultation on its automated decision-making guidance closed in May 2026, with final guidance now expected in winter 2026. The EHRC, for its part, has told Parliament directly that it sees a widening gap between how fast AI is being adopted and how ready the legal and regulatory system is to respond to it.
The realistic expectation is not a sudden flood of tribunal rulings. It is a steady accumulation of regulatory guidance, enforcement action, and the occasional test case, each one narrowing the room for employers to treat “the algorithm decided” as an adequate answer. Organisations that get ahead of that shift, by building the human oversight and transparency regulators are already asking for, will be dealing with policy and process now, rather than a tribunal claim later.
How Avado Can Help
Most AI-related workplace disputes start the same way as any other grievance: with a manager decision, or a manager’s response to a decision an employee didn’t understand or couldn’t challenge. Avado’s HR Compliance for Managers course, presented by employment law specialist Amanda Chadwick, builds the day-to-day judgement that prevents these situations from reaching a grievance or a tribunal in the first place, covering disciplinary and grievance essentials alongside the wider legal landscape managers now operate in.
Explore HR Compliance for Managers and equip managers to make the decisions that keep AI-related disputes out of the tribunal!
