Noyb files five privacy complaints over Scale AI’s Outlier worker data

Privacy organization noyb said on [October 27](https://noyb.eu/en/ai‑training‑platform‑outlier‑illegally‑surveils‑european‑platform‑workers) that it had filed five complaints against Scale AI in France, Italy, the Netherlands and Austria over practices on its Outlier training platform. The filings concern worker tracking, biometric identity checks and responses to requests for access to personal data. Outlier connects contributors with work such as writing prompts and ranking model answers. The complaints put the data collected from those contributors at issue, alongside the work they perform to improve AI systems. A person may be supplying training expertise while also supplying identity, device and activity information to the platform. These are allegations in complaints announced by noyb, rather than findings issued by a regulator. The group is asking the relevant authorities to investigate and impose a fine. The announcement does not establish that an authority has reached a decision.

In its [statement](https://noyb.eu/en/ai‑training‑platform‑outlier‑illegally‑surveils‑european‑platform‑workers), noyb alleges that trackers on Outlier send personal data to third parties while contributors are working. It lists IP addresses, contact details, unique identifiers and information about visited pages among the data involved. The group also challenges repeated biometric verification and alleges that access requests were not answered. Those claims raise different questions. Tracking concerns information generated while someone uses the platform; identity verification concerns the documents, images or recordings used to establish who the contributor is. The announcement does not, by itself, establish the complete technical scope of either practice. OpenTools has reviewed the complainant’s statement, but has not independently reproduced its tracking analysis or reviewed the full underlying research methodology.

Outlier’s [privacy policy](https://outlier.ai/legal/privacy‑policy), effective March 13, 2026, says Scale and providers such as Persona verify identities and skills to protect the service and reduce fraud. It describes collecting images and recordings, including selfies and government‑issued identification, and using them to generate facial measurements that may be biometric or sensitive data in some jurisdictions. This is an existing policy, not a verified response to the October 27 complaints. The policy says an alternative verification method may be available under certain circumstances. It also says Scale does not sell, lease, trade or profit from its defined “Verification Data,” which may be retained for up to two years after a person’s last interaction with the service, or as required by law. That retention statement concerns the policy’s defined Verification Data; other identification images and recordings have a separately referenced retention provision. The stated purpose and conditional alternative provide the company’s published account of verification. They do not resolve the complaint about whether the particular collection and tracking practices are justified. The next substantive development would be an authority’s procedural response or decision, or a specific company response addressing the filings. None is established by the two retained sources. *Lisa Steinfeld in a noyb press‑library portrait. Photo: Franziska Eilmsteiner / [noyb media center](https://noyb.eu/en/media‑center), supplied for reporting about the organization with photographer credit. Archive portrait; date not specified. This is not an Outlier worker, a biometric‑verification image or a photograph of the filing. Native crop and downscale; no artificial upscaling.*