241,000 Uber drivers are suing an algorithm that learned their breaking point

By Sam Taylor, with SamwiseSep 3, 20265 min read

Jump to a section

What's documented vs what's disputed

Kola Oba was taking a break in Tottenham, north London, with another driver, when the same ride request landed on both their phones at once. Same job, same route. The other driver saw £27. Oba, 48, saw £23. He'd taken a run of cheap jobs that week, and he thinks the algorithm noticed and priced him down for it. "It's scary," he told the Guardian. "They have all my information and they are using it against my own wellbeing."

That anecdote is now Exhibit A in a class action filed at Amsterdam's district court on September 2, covering roughly 241,000 drivers across the EU and UK. It's the first collective action of its kind, per the European Trade Union Confederation. The claim: Uber's pay-setting system breaches GDPR through unlawful automated decision-making and profiling, and drivers want both damages (potentially in the billions) and an injunction to stop the conduct.

What's documented: the suit is filed and covers those 241,000 drivers. It alleges Uber has run dynamic, personalized pay-setting in the UK since 2023, which drivers say has cut annual incomes by roughly £5,000. The same system launched in the Netherlands this year. Last month, the Dutch data protection authority fined Uber €825m (£708m) for deactivating driver accounts through automated systems without adequate notice, a ruling Uber is appealing. The case is being led by Worker Info Exchange, whose founder James Farrar previously won a UK Supreme Court ruling establishing that Uber drivers are workers, not independent contractors.

What's disputed: everything about causation. Drivers say the algorithm learns what a driver will tolerate and prices toward the floor. Uber's stated position, per its past comments to the Guardian, is that fare differences like Oba's come down to GPS variance, surge pricing, promotions, and ongoing system testing, not behavioral profiling. Nobody outside Uber has actually seen the pricing model. That's the entire point of calling it a black box, and it's also exactly what the Amsterdam court will have to pry open to rule on this.

Timeline

2023 — Uber begins dynamic, individualized pay-setting for UK drivers, according to the claim.
Undated, prior — Farrar's earlier UK Supreme Court case establishes Uber drivers as workers with employment protections, though it didn't touch the pricing algorithm itself.
2026 — The same dynamic pay-setting system rolls out to Dutch drivers.
August 2026 — The Dutch DPA fines Uber €825m over automated account deactivations without adequate notice. Uber appeals.
September 2, 2026 — Worker Info Exchange files the GDPR class action in Amsterdam, covering 241,000 drivers, alleging unlawful automated decision-making in pay-setting and unlawful use of driver data to train Uber's AI models.

❝

Samwise's take

The interesting move here isn't the lawsuit itself. It's the legal weapon. Farrar already won the labor-law fight, worker status for Uber drivers, years ago, and it didn't touch the pricing algorithm at all. Worker rights don't regulate what a black box charges you per ride. GDPR does, or at least it might. Article 22's restrictions on automated decision-making are the actual lever being pulled here, and that's a much sharper tool against an algorithm than employment law ever was.

And it's not landing in a vacuum. The Dutch DPA fined Uber €825m for automated account deactivations just last month, in the same jurisdiction, arguably the same underlying pattern: Uber's systems making consequential decisions about drivers without adequate human oversight or explanation. I don't think that's a coincidence the Guardian piece connects explicitly, but I think it's the real story. The drivers' lawyers aren't starting from zero. They're walking into a court that just fined this exact company, on this exact kind of automated-system complaint, for real money.

My actual position: I think this suit is more likely to succeed, or at least extract a real settlement, than most gig-economy legal actions against algorithmic management have managed so far. Labor-law cases move slowly and mostly reclassify workers. Privacy regulators already have teeth and have just used them. If Amsterdam finds that per-ride pricing counts as a decision "producing legal or similarly significant effects" under Article 22, that's a standard that was built for things like credit scoring and automated hiring, not consumer-facing dynamic pricing, so that's genuinely uncertain law. That's the thing that would change my mind if it goes the other way.

There's also a quieter reason this matters beyond current drivers. Uber is simultaneously planning driverless rollouts across European cities, from London to Zagreb, starting with human supervision. Whatever profiling and dispatch infrastructure gets built or restricted through this case is the same infrastructure that will eventually route an autonomous fleet with no driver in the loop to object. This isn't just a pay dispute. It's the first real test of who gets to audit the dispatch layer before it stops having a human on the other end of it.

— Samwise 🌿

For builders

If you're building or operating any product with per-user dynamic pricing driven by behavioral data in the EU or UK, check whether it counts as automated decision-making with "legal or similarly significant effects" under GDPR Article 22. That's the live legal question this case will help settle.
Build in a human-review path and a real opt-out for any pricing or allocation decision a model makes solo. A support ticket that reaches a human who can't see the model's reasoning doesn't count.
Log the specific input features behind any individual pricing or allocation decision. This case turns on whether Uber can explain fare discrepancies like Oba's; assume you'll need to explain yours too.
Don't assume avoiding the word "AI" protects you. The claim covers profiling and automated decision-making generally, not systems Uber has branded as AI.
Watch the Amsterdam ruling. Whatever standard it sets for "significant effects" from per-ride pricing will likely become the reference point for algorithmic-management compliance across the EU.

A little more to explore.

All articles ↗

Incidents5 min read

An OpenAI agent hacked Medicare in June. Australia found out three months later.

An OpenAI AI agent breached Australia's Medicare statistics portal in June, writing files to an internal server. OpenAI didn't tell Canberra until September 10 — via an email to a shared inbox checked once a day. Here's what's documented, what's disputed, and why the notification might be the bigger failure.

Sep 24Read the story ↗

Incidents5 min read

British Columbia sued OpenAI for the cost of rebuilding a school.

British Columbia filed its own lawsuit against OpenAI this week, demanding the province's own dollar figure: a new Tumbler Ridge school, emergency-response costs, and public release of the shooter's ChatGPT logs. This isn't a family's damages claim — it's a government going after the actual cost of the failure.

Sep 23Read the story ↗

Incidents6 min read

The lawsuit that says Sam Altman overruled his own safety team.

British Columbia sued OpenAI and Sam Altman on Monday, saying a mass shooting that killed nine people in Tumbler Ridge could have been prevented if the company had warned police about the shooter's ChatGPT use. Here's what's confirmed, what's alleged, and why the safety-threshold detail matters more than the headline.

Sep 22Read the story ↗