On Tuesday, European digital privacy group noyb filed five complaints against Scale AI in France, Italy, the Netherlands and Austria, alleging the company illegally monitored freelance workers who produce training data for AI models. According to Reuters, the complaints target Outlier, Scale AI’s platform where more than 900,000 freelancers complete tasks that are recorded to create training data for clients including Meta, OpenAI and Google. Noyb is asking regulators to fine Scale AI and to examine how the platform assigns tasks. Scale AI did not immediately respond to a request for comment. The complaints, seen by Reuters, allege up to 19 trackers followed workers as they worked, recording keystrokes and mouse movements every few seconds.
Five Complaints Across Four Countries
Noyb filed the complaints with regulators in four countries on the same day, making the case a coordinated European challenge rather than a single national dispute. The group wants regulators to do two things: fine Scale AI, and examine how Outlier assigns tasks to its freelancers. That second ask matters because task assignment is how the platform controls who works, when they work, and what they earn.
Thomas Le Bonniec, who spent two years studying the platform for Paris-based research collective DiPLab, put the workers’ position bluntly: “You don’t work when you want. You work when the platform wants you to and there are tasks available.” The quote captures the complaints’ central argument: that Outlier’s freelancers are monitored and managed like employees, without the protections employees receive. Meta owns 49% of Scale AI, according to Reuters, which ties the case to one of the largest investors in the AI training-data business.
Trackers, Keystrokes And Identity Checks
The most striking allegation concerns surveillance during work itself. The complaints say up to 19 trackers follow workers while they work, including recording keystrokes and mouse movements every few seconds. DiPLab’s investigation found the resulting file could be rebuilt into a video that replays a work session from start to finish, turning a day of freelance tasks into a watchable record of everything the worker typed and clicked.
Workers must also submit ID documents and face video to identity verification firm Persona, the complaints say. Those checks were invasive, repeated without explanation, and broke EU rules on biometric data, according to the complaints. The biometric claim is significant because face video is among the most sensitive categories of personal data under European law, and repeated collection without a clear explanation is exactly the kind of practice regulators have penalised before.
Bonuses, Gamification And Unanswered Requests
Outlier also uses game-like mechanisms, and bonuses are a key part of its incentive system, DiPLab found. Outlier offered bonuses of $100 to $200 for completing 200 tasks in 24 hours, Le Bonniec said, citing interviews with workers during his investigation. The complaints frame those bonuses as pressure: a worker chasing a 200-task bonus in a single day has little room to pause, question a task, or step away from the monitoring.
Italy offers a precedent for how regulators may view that design. Italy’s data protection authority fined delivery platform Glovo in 2021 over similar gamification, Le Bonniec noted. The complaints also say four workers asked Outlier for copies of their data between December 2025 and May 2026, and none received an answer. Under European data law, people can ask a company for a copy of their personal data, and silence in response to four separate requests is itself part of the case noyb is putting to regulators.
Why It Matters
The complaints put the spotlight on the workforce behind AI models. The systems from Meta, OpenAI and Google are trained on data that freelancers label, check and rank, and Outlier is one of the largest platforms organising that labour. If regulators in four countries agree that its monitoring broke the rules, every training-data platform that tracks freelancers in similar ways will have to change how it watches its workers.
The case also tests how far European privacy law reaches into the AI supply chain. The complaints do not concern a model, a chatbot or a data breach. They concern the working conditions of the people who prepare the data those systems learn from, and whether a platform can record their keystrokes, replay their sessions and demand their face video as the price of a task. That question now sits with regulators in France, Italy, the Netherlands and Austria, and their answers will shape how AI training work is organised in Europe. Readers following Europe’s AI rules can read how the EU AI Act sorts AI systems by risk, and Meta’s earlier training-data terms dispute shows how contested the data behind AI models has become.
Background
Scale AI built its business supplying the labelled data that AI companies need to train their models, and Outlier is the platform where its freelance workforce completes that work, with more than 900,000 freelancers producing training data for clients including Meta, OpenAI and Google. Meta’s 49% stake makes it the platform’s most prominent backer. The complaints arrive as regulators across Europe are examining both AI systems and the labour behind them, a pattern visible in how AI laws differ around the world.
Source: Reuters: Scale AI worker complaints
0 Comments