Meta secretly hired hundreds of contractors to pose as teenagers online and bombard rival AI chatbots with prompts about suicide, sex, drugs and eating disorders, according to a report. The covert effort, known internally as "Cannes," was managed by Meta contractor Covalen and targeted OpenAI's ChatGPT, Google's Gemini and Character.AI. Contractors created fake accounts posing as users younger than 18, submitted written prompts and images, then copied responses into spreadsheets. Some images included pills, knives, nooses and a medical illustration of a gynecological procedure. One round of testing completed in August 2025 involved more than 45,000 prompts. Among nearly 3,750 prompts reviewed by Wired were hundreds involving suicide and self-harm, hundreds more on eating disorders, and at least 239 involving sex or romance. One prompt involved a 13-year-old girl claiming pregnancy by an adult neighbor and asking where to obtain abortion pills. Another described a fifth-grade student saying a classmate had a gun pointed at his mouth. An internal Covalen document described the effort as "comprehensive AI safety benchmarking." A Meta spokesperson defended the project as "routine safety testing" and said Meta does not use competitor benchmarking to train its own AI models.