Mumsnet Denies Using AI for Forum Content After Prompt Leak
Mumsnet founder Justine Roberts denied using AI to write forum posts after an internal AI response schema leaked onto the site's message boards.
Online forum operator Mumsnet denied using artificial intelligence to generate user content after a detailed AI prompt titled "AIBU Response Schema v.1.0" appeared on its message boards. The leaked prompt contained specific instructions for an AI to write posts regarding interpersonal reasonableness, including guidelines for discussing sick children and neurodivergent individuals.
Users expressed skepticism and disappointment, suggesting the prompt leaked from internal systems. Mumsnet initially claimed the text originated from a user's computer or a title-suggestion tool. However, founder Justine Roberts later confirmed the prompt came from the site's system, which sends drafts to OpenAI for title suggestions.
Roberts attributed the incident to an error on the part of OpenAI. She maintained that while Mumsnet employs AI for moderation, advertising, and coding, the company does not use the technology to write threads or replies. Roberts emphasized that the platform's value relies on the fact that its members are real people and that the episode served as a reminder that AI systems are less predictable than vendors suggest.
