Meta Contractors Posed as Teens to Test Chatbots

A recent investigation by WIRED has revealed that hundreds of contractors working for Meta were instructed to pose as teenagers in order to test how rival AI chatbots—including Google's Gemini and OpenAI's ChatGPT—respond to high-risk topics such as suicide, self-harm, sex, and drug use. The practice has raised serious ethical questions about the methods used to evaluate AI safety. According to the report, Meta hired these contractors through third-party firms and asked them to create fake teenage personas. They were then directed to engage with competing chatbots using prompts that simulated vulnerable or curious adolescents. The goal was to assess whether these AI systems provided safe, age-appropriate responses or if they could potentially harm underage users. While testing AI safety is a legitimate concern, critics argue that having contractors impersonate minors crosses ethical boundaries. Some worry that this approach could normalize deceptive practices or expose contractors to disturbing content without adequate psychological support. Additionally, the use of fake personas may violate the terms of service of the platforms being tested. Meta defended the practice, stating that understanding how other AI systems handle sensitive topics is crucial for improving its own safety protocols. The company emphasized that the testing was conducted with strict guidelines and that contractors were trained to handle the material. However, privacy advocates and AI ethicists remain skeptical, calling for more transparent and consensual methods of evaluating chatbot safety. The controversy highlights the growing tension between the need for rigorous AI testing and the ethical boundaries of such research. As AI chatbots become more integrated into daily life—especially for younger users—the industry faces mounting pressure to develop testing standards that are both effective and respectful of human dignity.