Add the world of technology: follow us on Google News

A report from Wired magazine revealed that Meta conducted a secret program in which it hired hundreds of contractors to impersonate teenagers, and they began sending shocking inquiries and requests to the AI systems of its competitors, such as ChatGPT, Gemini, and Character.AI, using fake accounts claiming to be users under 18 years old.

According to the report, these inquiries included sensitive topics such as suicide, eating disorders, and drugs, as well as sexual themes or romances, all written in the voice of children and teenagers, along with images of sharp objects and medications.

The documents indicated that this was part of Meta's intensive 'AI safety testing' process, but it did not inform its competitors nor share the test results for transparency.

Meta launches Meta Glasses AI glasses starting at $299

Meta described this procedure as standard practice for the sake of safety, while organizations specialized in AI ethics considered that concealing the experiment from other parties and burdening employees with such psychological loads raises questions about the ethics and mechanisms of controlling these experiments.

A number of contractors expressed shock and the harshness of what was asked of them, which brings back to the forefront the ongoing issue of the impact of this type of work on the mental health of digital technology sector employees.

Tags

Meta artificial intelligence

Editorial team: follow on X, send email, last updated: July 4, 2026

One minute