The pay algorithm has become a workplace dispute. Food delivery riders and Uber drivers want platforms to explain how computer systems decide which jobs they receive and how much they earn. The demand targets the “black box” at the centre of gig work—software that manages labour while revealing little about its decisions.
In Edinburgh, a group of food delivery riders says pay rates have fallen and working conditions have deteriorated as platforms increased their use of automation. David, a food delivery rider, said: “I am making half the money I was making four years ago, for the same amount of hours. It makes no sense.”
Xabier Villares, a Deliveroo rider and lead organiser of the Workers’ Observatory, said: “There has been a dramatic change in the last three years.” The observatory is a charity founded by gig economy workers alongside academics at St Andrews and Edinburgh universities, and it has focused on finding out how platforms organise work.
The Workers’ Observatory has run experiments tracking the different pay levels algorithms offer workers. The practice, known as “dynamic pricing”, has become a target for trades unions, which are campaigning to ban it because workers cannot predict their earnings from one shift to the next.
Research from the University of Oxford and Columbia business school found that Uber drivers earned “substantially less” an hour after the company introduced a dynamic pricing algorithm in 2023. That finding gives the dispute a harder edge: the system is not only difficult to inspect, but may also change the value of the work itself.
The hidden machinery behind a shift
Cailean Gallagher, director of the Workers’ Observatory and a lecturer at St Andrews university business school, said the group wants to understand the systems before demanding changes to isolated features. “What we are trying to do with the observatory is to get on the ground floor in terms of understanding how the whole apparatus works,” he said.
Gallagher added that “there is so much infrastructure of knowledge and data that’s concealed” by platforms. That data can include customer ratings, accepted or canceled orders, service times, location, times and other measures used to organise work and evaluate workers.
The same problem has reached the courts. Drivers from the UK, the Netherlands and other countries launched a landmark class action against Uber, claiming they live in “constant fear” of the “soulless” algorithm used to set pay and allocate jobs.
Mohammed Shirwa, an Uber driver from Rotterdam, claims Uber’s pay-setting mechanism preys on weakness. The lawsuit says Uber’s system breached privacy laws and pushed down earnings, while Uber denies adjusting trip prices based on an individual driver’s behaviour and attributes discrepancies to features such as GPS.
Transparency is not the same as an explanation
The dispute is not limited to the UK or Uber. In Latin America, delivery and transportation apps measure location, times and ratings, but workers do not always know how those data influence their work and income.
Derechos Digitales, an organisation defending digital rights in Latin America, has identified opacity in the way data shapes decisions. It has called for transparency and explainability in algorithmic decision-making—two terms that sound interchangeable until a worker needs to challenge a lost shift.
Transparency means knowing which elements a system uses. Explainability means understanding why those elements produced a particular decision. Mexico’s Ley Federal del Trabajo requires platforms to inform workers about the factors influencing their tasks and earnings in clear language.
A 2024 study by the Universidad Nacional de General Sarmiento analysed 70 interviews and 750 cases involving delivery workers in Buenos Aires. It found that metrics such as ratings and acceptance rates influence job opportunities and earnings, while safety concerns and caregiving responsibilities can affect acceptance rates and then reduce income.
Research from Yale University and the University of Singapore found that racial bias in client ratings affected worker evaluations and income. Discrimination increased by 33.9% the proportion of lower ratings among workers perceived as minorities and produced a 6.5% income gap.
Mexico, Colombia and Uruguay recognise the importance of human review when automated systems make decisions. Review and supervision are meant to ensure that decisions affecting workers remain reasonable and proportional, rather than standardised, opaque and discriminatory.
That safeguard raises its own questions. Workers still need to know how long reviews should take, whether reviewers are independent and who provides external oversight. Delays can leave people without access to work for days or weeks, cutting off income while an invisible process considers their case.
Even if a platform lists the variables its system uses, one question remains: who verifies that the system functions as claimed? Until workers can answer that, “transparency” risks becoming another interface layer over the same old black box.
Based on
- Food delivery riders call on platforms to open up AI ‘black box’ they say has cut pay — theguardian.com
- Uber drivers launch European class action over ‘soulless’ and ‘scary’ AI algorithm | Uber | The Guardian — theguardian.com
- Mi “jefe” es un algoritmo: cómo la automatización afecta a los repartidores en América Latina | WIRED — es.wired.com
