Skip to content
Project

SHIFT-AI (Safeguarding Human-AI Interaction via Feedback from Teens)

SHIFT-AI Team Picture

How can we make generative AI safer for young people?

Welcome to SHIFT-AI. This project, supported by the UK Government’s AI Security Institute (AISI) Challenge Fund, brings young people to the centre of conversations about how generative AI can be made safer for teenagers and young adults.

Based at the University of Oxford (Department of Psychiatry) and developed within the Design Bioethics Lab, SHIFT-AI works directly with young people to explore how generative AI may be used by adolescents, what risks may arise, and how these systems can be made safer for young users.

Our project brings young people in as co-researchers, working with us to identify risks, design safeguards, and develop guidance for developers and policymakers. This is central to our co-production approach.

Why co-production?

Most research on AI safety focuses on adult users, technical risks, or hypothetical harms. However, young people often use generative AI in ways influenced by their developmental context - how they learn, relate to others, and make decisions. These patterns are difficult to capture through adult-led research alone.

Across five co-production sessions with the NeurOX Young People’s Advisory Group, we have:

  • mapped how generative AI may be used in everyday adolescent contexts 
  • identified risks that feel realistic and relevant to young people 
  • explored what makes adolescents particularly vulnerable to these risks 
  • co-designed and tested potential safeguards for young users 

This approach allows us to identify risks and solutions that may not be visible in surveys, experiments, or purely technical analyses. It also ensures that the outputs of the project - such as guidance for developers, policymakers, and young users themselves - are grounded in real-world use and informed by those most affected.

Project outputs

We’re actively developing a set of youth-led outputs and will share them here as they’re released. Bookmark this page to stay updated or follow our team on LinkedIn. Here’s a preview of what’s to come:

  • The SHIFT-AI Youth AI Safety Toolkit
    We’ll be releasing a series of youth-led tools for using generative AI safely, including practical advice, examples, and strategies for navigating common risks.
  • Policy recommendations
    We’ve created a set of evidence-based insights to support UK policy for young users of generative AI.
  • Public engagement and media
    Early insights from the project have contributed to public debate on youth and AI safety, including a letter published in The Guardian highlighting the need to extend child online safety policy to include generative AI use and its developmental impacts.