When a medicine causes an unexpected effect, there is somewhere to report it, and those reports accumulate into evidence. When a conversational AI tool or a social media platform leaves someone feeling worse - more anxious, more isolated, unable to stop - there is usually nowhere for that to go. It stays a private experience, and nothing aggregates.
We are launching a public reporting platform to change that. It collects reports of distress that people notice during or after using conversational AI tools and social media, so that patterns of harm can be studied rather than guessed at. Reporting is anonymous by default, needs no account, and asks for no name.
The platform is part of a wider effort: the European Commission's living guidelines for the responsible use of generative AI in research, and the research behind them - Living guidelines for generative AI and ChatGPT: five priorities for research.
The platform launches soon. When it is live, the report form will be at /report.