Shaping responsible innovation through AI: Evidence from UKRI-funded social science projects

Artificial intelligence (AI) is rapidly changing how social science research is designed and conducted, opening new possibilities for analysing complex data, generating evidence and addressing societal challenges. But as AI becomes more deeply embedded in publicly funded research, important questions arise about how these technologies are being used and whether principles such as transparency, accountability, fairness, privacy and responsible innovation are keeping pace. This report examines 3,287 AI-related UKRI-funded social science projects to provide a large-scale picture of how AI methods have evolved across the UK research portfolio. It finds a diverse and increasingly layered landscape: machine learning remains the dominant approach, established statistical and computational methods continue to play an important role, while natural language processing, deep learning and, more recently, foundation and generative AI are expanding rapidly across different disciplines.

The findings also reveal an important challenge for research funders and policymakers. Responsible AI concerns are becoming more visible, particularly around privacy and security, governance, transparency, trust, fairness and ethics, but they remain unevenly embedded across AI-enabled projects. The report therefore proposes a practical framework for UKRI to evaluate responsible AI throughout the research lifecycle, structured around six principles: anticipation, reflexivity, inclusivity, responsiveness, transparency and equitability. By connecting evidence on emerging AI capabilities with a framework for responsible research practice, the report offers a basis for monitoring how AI is transforming research, strengthening governance and ensuring that future AI-enabled research is not only innovative, but also transparent, trustworthy and socially beneficial.

Download the full report using the button on the right