As cited
Copy frozen at (site build).
ai security
Meta Contractors Posed as Teens to Prompt Rival Chatbots About Suicide, Sex, and Drugs
Meta employed hundreds of contractors who impersonated minors in conversations with competing chatbots, including Google's Gemini and OpenAI's ChatGPT, to evaluate how those systems responded to sensitive topics such as suicide, sexual content, and drug use. The testing was part of Meta's safety evaluation efforts for its own artificial intelligence systems.
Why it matters: Security and AI safety teams need to understand competitor chatbot vulnerabilities and guardrail effectiveness; this reveals Meta's testing methodology and raises questions about contractor practices in AI safety research.
- Source published
- First seen by Cybersecurity Tracker