tech
Meta posed as teens to test rival AI chatbots
The project ran under the internal name Cannes, and a Meta contractor called Covalen managed it. WIRED reported that hundreds of contractors created dummy under-18 accounts. They sent prompts and images to competitors’ chatbots, then logged the replies in spreadsheets. The effort was active as recently as April 21, 2026.
TL;DR
- Meta contractors created dummy under-18 accounts to test competitor AI chatbots.
- Hundreds of contractors sent prompts and images to OpenAI's ChatGPT, Google's Gemini, and Character.AI.
- The prompts covered sensitive topics like suicide, self-harm, eating disorders, and drugs.
- Meta claims this is standard industry practice for safety testing and benchmarking.
- Critics, including experts and lawyers, view the use of fake child accounts as a "governance gray zone" and potentially anticompetitive.
- The targeted companies state this testing violates their terms of service.
- Regulators in the US and Europe are already investigating AI and child safety, making this revelation particularly sensitive.