A privacy group filed complaints against Scale AI over alleged surveillance of its workers

The European privacy group noyb filed five complaints in Europe on October 27 against Scale AI over how its training platform Outlier treats its workers. Euronews reports two complaints in Austria and one each in France, the Netherlands and Italy. Noyb is asking regulators to fine Scale AI and to examine how the platform assigns work. These are allegations in complaints, not findings by regulators.

Outlier recruits freelancers and subject-matter experts to help train, evaluate and improve AI models. Noyb says it uses more than 900,000 of them, with customers including Meta, OpenAI, Google and the US government. The Times of India, citing Reuters, says the complaints allege up to 19 trackers monitor workers, including by recording keystrokes and mouse movements every few seconds. Noyb adds that trackers from companies such as TikTok, LinkedIn and Facebook are placed on workers' devices, with access to data like IP addresses and contact details.

Euronews also reports that workers must repeatedly go through biometric verification run by an identity company called Persona, which involves a video recording of their face and an uploaded ID. According to the complaints, four workers asked Outlier for copies of their personal data between December 2025 and May 2026, and none received an answer. Lisa Steinfeld Leiser, a data protection lawyer at noyb, asked: "Why should LinkedIn or Facebook know which tasks you're working on?"

A research collective, DiPLab, found that the resulting activity data could be reconstructed into a video replaying a work session, according to the Times of India. Its researcher Thomas Le Bonniec also questioned how much control freelancers have over their schedules, and cited bonuses of $100 to $200 for completing 200 tasks within 24 hours. Inshorts notes that Meta owns 49% of Scale AI.

Why it matters for AI builders: the human labeling and evaluation layer behind many models is now drawing privacy scrutiny in Europe, at a moment when the EU's AI Office has opened investigations into OpenAI and Anthropic, Euronews reports. If your team buys annotation or evaluation work from a vendor, ask what is tracked on contractors' devices, how biometric checks are handled, and how data access requests are answered.

What happens next depends on the data protection authorities in the four countries, which now decide whether to investigate. Scale AI's response was not part of the material we reviewed.

Building with AI? See this week's hot AI repos and how to use each one.