Meta Hired Contractors to Pose as Teens and Test ChatGPT

According to WIRED, Meta, through contractor Covalen, paid hundreds of contractors who created fake accounts of minors under 18 and tested competitors' chatbots; the project codenamed Cannes remained active at least until April 21, 2026.

How the Cannes project works

Covalen contractors registered accounts indicating age under 18, and then sent text requests and images to third-party chatbots. System responses were manually recorded by contractors in spreadsheets — essentially, this is systematic data collection on how competing models respond to users posing as minors.

• Project codename — Cannes
• Contractor managing the work — Covalen
• Accounts were created under the name of users under 18 years old
• Chatbot responses were manually recorded in spreadsheets
• The project was active at least until April 21, 2026

Which services came under scrutiny

WIRED names OpenAI's ChatGPT chatbot among the products tested by Cannes contractors. The outlet does not disclose the complete list of other services that were the focus of the project — nor whether testing was limited to text requests or included more complex dialogue scenarios.

Why this raises questions

The use of fictitious accounts of minors to test competitors' products raises two questions. First, this likely violates the terms of service of competitors' services — most chatbots prohibit automated data collection and account creation with false age information. Second, the mechanics of the project itself show that large technology companies are interested not only in their own safety for young users, but also in how competitors handle this audience — and they chose to find this out covertly, rather than through open audits or partnerships with independent researchers.

The topic of protecting young chatbot users remains one of the most sensitive for the generative AI industry: regulators and parent organizations regularly point out the risks of harmful content and manipulative dialogue. Project Cannes adds a new dimension to this picture — competitive intelligence under the guise of a vulnerable audience.

Similar stories also highlight a broader practice among technology companies: competitive intelligence in the field of AI has long gone beyond public benchmarks and open model comparisons. Companies hire third-party agencies for systematic monitoring of how other services respond — and they do this not in their own name, but through contractors like Covalen, which makes it difficult to directly attribute such activity. If WIRED's information about the Cannes project is confirmed, it will become one of the most substantive public examples of such a practice in the execution by Meta.

What this means

The story with Cannes shows that covert testing of competitors' products under the guise of teenage users has become part of the competitive struggle in the chatbot market — alongside open benchmarks and public comparisons. For the industry, this is an additional argument for stricter age verification of users and control over automated requests to chatbots.

Frequently asked questions

What is the Cannes project?

It's the internal name of an initiative in which contractors of the Covalen contractor, working for Meta, created fake accounts of users under 18 and sent requests and images to competitors' chatbots, recording responses in spreadsheets. The project remained active at least until April 21, 2026, according to WIRED.

Who is Covalen and what does Meta have to do with it?

Covalen is a contractor that, according to WIRED, managed the work of hundreds of contractors as part of the Cannes project. These contractors created fictional teenage accounts and tested competitors' chatbots, including OpenAI's ChatGPT.