DYNAMIC PRICING: TUC report warns algorithmically set pay is driving private hire and courier workers into poverty and ill health
- Perry Richardson
- 3 days ago
- 4 min read
Platform operators including Uber and Deliveroo are deploying algorithmic wage-setting systems that have materially reduced take-home pay for private hire drivers and food delivery riders across the United Kingdom, according to a report published by the Trades Union Congress.
The research, which includes testimony from 11 platform workers and draws on a 2025 University of Oxford study analysing 1.5 million Uber trips made between 2016 and 2024, argues that so-called dynamic pay constitutes a new and harmful wage-setting regime that warrants urgent government action.
Dynamic pay, as defined in the ’The Human Price of Dynamic Pay’ report, refers to algorithmically determined, variable, and potentially personalised worker compensation for a task. Under this model, two workers completing an identical job may be paid entirely different rates, and an individual worker performing the same task on consecutive days may receive vastly different amounts. At the moment a job offer appears on screen, workers often cannot determine what they will earn for accepting it.
The Oxford study, conducted in collaboration with Worker Info Exchange, found that following Uber’s introduction of dynamic pricing in 2023, its surplus per hour increased by 38 percent while drivers’ real pay fell. The median driver retained just 71 percent of the fare, and only 46 percent of drivers retained 75 percent or more. Where Uber had long stated a nominal 25 per cent commission, dynamic pricing transformed this into a variable and personalised deduction, with drivers on many trips retaining only 50 to 60 percent of what passengers paid. Unpaid waiting time also increased substantially, with separate research finding riders and drivers spend an average of ten hours per week logged on to apps without generating any income.
The human cost documented in the report’s case studies is considerable. Mehmet, a London-based Uber driver of nearly nine years, suffered a heart attack following a slow day on the platform. He attributes the event in part to the sustained psychological pressure generated by unpredictable earnings and the platform’s six-second job acceptance window, which forces near-instant decisions on financial viability while driving. Kola, another London driver with more than eight years on the platform, describes working 12-to-15-hour days to compensate for falling rates while carrying responsibility for four children and an elderly mother. “You are working, working, working,” he told researchers, “and you are not even earning enough. It is a cycle, the same devil cycle.” Maryan, a single mother and primary carer for a special-needs parent, now works 50 to 60 hours a week compared with the five or six hours she once managed before dynamic pay was introduced.
The report identifies a set of structural conditions that have enabled dynamic pay to take hold. Deliberate oversupply of labour gives platforms leverage to keep rates low, as tens of thousands of workers compete for a limited volume of jobs.
Workers are typically not compensated for waiting time, meaning they absorb the financial risk when demand is slower than anticipated. Employment status ambiguity further compounds the problem: platforms frequently classify workers as self-employed, stripping them of minimum wage protections and collective bargaining rights, even as algorithmic systems exert detailed managerial control over their working patterns. The TUC notes that enforcement of employment and data rights by state bodies remains weak, and that the atomised, geographically dispersed nature of platform workforces limits the effectiveness of trade union organisation.
Inequality between workers has widened under dynamic pay. The Oxford study found that, after Uber introduced the system, 93 out of 114 drivers in the sample earned less, while only 21 earned more. Earnings predictability effectively collapsed: researchers used machine learning to demonstrate that pay for similar trips could no longer be forecast from historical patterns. The report also draws attention to gendered effects.
Women working on ride-hail platforms frequently avoid night shifts for safety reasons, but those shifts carry the highest surge rates. Zara, a Midlands-based driver who uses a pseudonym in the report, told researchers she earns “the lowest fares” during daytime hours when she feels safe to work. “Throughout the day, they’ll give us the lowest fares,” she said. “It’s just absolutely horrendous prices.”
Worker Info Exchange has initiated collective legal action challenging Uber’s use of dynamic pay on behalf of drivers in the UK, the Netherlands, and elsewhere in Europe, arguing that the algorithmic wage-setting process is non-transparent and breaches drivers’ data rights. The action seeks both transparency and financial redress for lost income.
In parallel, the TUC is calling on the Government to take three legislative steps before the end of the current parliament in 2029: ending the use of dynamic pay altogether and returning to a model based on transparent time-and-distance rates with consistent platform commissions; reforming employment status to bring more platform workers within the scope of worker protections and ending what the report describes as bogus self-employment; and introducing collective data rights that would allow trade unions to access the information platforms use for algorithmic wage-setting and to bargain meaningfully over pay.
The Employment Rights Act 2025, which received parliamentary approval following the Government’s Plan to Make Work Pay, introduces guaranteed hours contracts for variable-hours workers and extends unfair dismissal protections, while also granting trade unions a right of digital access to workers. The TUC argues this legislation, while welcome, does not adequately address the specific dynamics of algorithmic pay-setting in platform labour markets. The report points to the European Union’s Platform Work Directive and AI Act as comparative frameworks, both of which impose transparency requirements and human oversight obligations on automated management systems. It also notes that New York State now requires businesses using personal consumer data in pricing algorithms to disclose this at the point of sale, though the TUC concludes that disclosure requirements alone would not protect platform workers given the structural power imbalance they operate within.
The Competition and Markets Authority stated in the summer of 2025 that it was more likely to be concerned about dynamic pricing where consumers face short decision windows, where vulnerable groups are disproportionately disadvantaged, or where pricing mechanisms reduce competitive market entry. The TUC argues these criteria apply directly to dynamic pay in labour markets, where workers face acceptance timers of five or six seconds and where those most financially dependent on the platform are most likely to accept below-viable rates.
