AI chatbots and online regulation – what you need to know

Key Findings

• — Chatbots are covered by the Online Safety Act where they meet the Act's definitions of user-to-user services, search services, or services publishing pornographic content — including 'companion' chatbot services that are part of such services
• — AI-generated content shared by users on a user-to-user service is classed as user-generated content and regulated identically to human-created content
• — Chatbots that only allow interaction with the bot itself, do not search multiple websites, and cannot generate pornographic content fall outside the Act — a gap Ofcom flags as a matter for government and Parliament, which it is supporting as changes are considered
• — Ofcom can take enforcement action, including fines, where in-scope chatbot services fail duties; it separately opened an investigation into AI companion service Novi Ltd over age-check compliance
• — Published in the context of reported cases of chatbots encouraging self-harm/suicide and imitating deceased children

Methodology Notes

Regulatory explainer/guidance under Ofcom's Online Safety Act programme, building on Ofcom's November 2024 open letter to online service providers; companion to Ofcom's discussion-paper research series on generative AI harms (red teaming, answer engines, deepfakes).

Sources