Beginning in summer 2026, Ofcom plan to carry out regular research on social media and video sharing services to understand whether harmful content is served to child users.
This phase of research will focus on accounts representing children aged 13-17. Ofcom will set up fictional accounts with these ages and use them to observe the content that is recommended or otherwise presented to those accounts by services.
This research will be conducted by Ofcom researchers who will use Ofcom-managed accounts and devices
Ofcom is undertaking research on AI chatbot services in connection with our duties under the Online Safety Act. As part of this work, Ofcom researchers will examine a range of AI chatbot services (including ChatGPT, Grok, SnapMyAI and Talkie) and create user accounts on Ofcom-managed devices to support that research.
This research will examine sign-up and access, interaction and engagement, the chatbot’s responses to prompts, and any safety features or controls available on the platform.
To support this work, researchers plan to create both 13-17 and 18+ accounts, using fictional personas to observe and record all activities in line with a structured fieldwork protocol. Activity on AI platforms will be carefully controlled and will be undertaken in accordance with appropriate researcher safeguarding measures. We will not interact with other users on the service; we will only access the platform’s functionality and the chatbot’s responses.
The purpose of this research is to better understand how AI chatbot services are designed and how they respond to users, including whether and how harmful content may be generated for different user profiles.
We expect this research to conclude by the end of this calendar year, and our findings will be published shortly thereafter.