Five independent workers have filed a complaint against the American company Scale AI in four European countries, including France. They denounce an excessive collection of their personal data, accusations that the company, contacted by franceinfo, disputes.
- Five workers filed complaints against Scale AI in four European countries.
- They accuse the company of collecting excessive personal data through trackers, screenshots and identity checks.
- Scale AI rejects the accusations and says contributors consent to these tools.
- Noyb asks authorities to investigate, but no decision has been made yet.
On Tuesday, October 27, the complaints were filed. They target Scale AI, specialized in training artificial intelligences, and were launched, according to Lisa Steinfeld Leiser, lawyer at the association noyb, “in France, Italy, the Netherlands and Austria”. Noyb, which defends digital rights, drafted these complaints.
At the heart of the case is Outlier, a platform that claims more than 900,000 contributors worldwide. These workers perform paid tasks, notably to evaluate responses generated by artificial intelligences, compare them or correct them. Scale AI counts among its clients Meta, Google and OpenAI.
Contested surveillance tools
The complainants denounce the use of digital surveillance devices, called “trackers”, that may collect information about their activity. Some would allow tracking their browsing when they leave the Outlier platform to visit other sites.
These practices have been studied for two years by researchers of the French collective DiPLab, who identified up to 19 trackers and various surveillance tools. Some could enable reconstructing a work session as a video. Antonio Casilli, researcher at Télécom Paris, describes these devices: “There are extensions that monitor the mouse movements of these workers, that take screenshots of their sessions or that can sometimes be used to fully reconstruct their activities, as if it were a video.”
The identity checks required to access certain tasks are also criticised. Workers must provide an identity document and facial images, sometimes repeatedly, with the involvement of a specialised provider for identity verification. For the complainants, these devices lead to excessive data collection. They also point out the lack of response to some requests for access to the information held about them.
The company’s response
Scale AI rejects the accusations, which it deems “baseless and inaccurate”. It nevertheless acknowledges that, for certain tasks, software may record the time spent on tasks and screen activity. According to the company, contributors are informed of these practices, must give their consent and activate the tool themselves.
According to Scale AI, these controls aim, among other things, to prevent fraud, protect its clients’ confidential information and ensure accurate payment of workers. Identity documents and facial images would be used to verify contributors’ identities and secure their accounts. The company also states that it complies with the General Data Protection Regulation (GDPR).
Noyb asks authorities to investigate Scale AI’s practices and consider sanctions. At this stage, the allegations made in the complaints have not been subject to any definitive decision by the competent authorities.
