Trust unpacked: Understanding trust and distrust through AI-enabled analysis

What is the issue?

Public trust in government and public institutions is essential for effective policymaking and democratic legitimacy. However, evidence suggests that trust in institutions may be declining across many democracies, including the UK. Existing approaches to measuring trust provide only a snapshot of public attitudes at a single point in time. They often fail to capture how trust and distrust change over time and what they mean in real-world interactions.

This project responds to these challenges by exploring new ways of understanding trust and distrust in practice. Rather than relying solely on traditional survey measures, the project will investigate whether social media data and advances in artificial intelligence can provide richer, dynamic insights into how people experience and express trust in public institutions.

The research will focus initially on health and criminal justice policy settings, helping to identify how trust and distrust emerge in different public-sector contexts and what this means for policymakers seeking to strengthen public confidence and engagement.

How are we helping?

The project will develop and test an innovative framework for analysing trust and distrust through a combination of behavioural science, computational social science and AI-enabled text analytics.

Over 12 months, the research team will:

• Develop a framework that distinguishes institutional trust from distrust and identifies how both can be observed in practice.
• Conduct case studies in health and criminal justice to understand how trust and distrust are expressed in different policy domains.
• Design and validate an AI-enabled analytical tool that combines natural language processing, sentiment analysis and behavioural indicators to assess trust and distrust using publicly available online data.
• Compare findings across sectors to identify similarities and differences in patterns of trust and distrust.
• Explore how the resulting insights can support the development of targeted trust-building interventions and more responsive policymaking.

By combining conceptual innovation with cutting-edge AI methods, the project aims to move beyond static measures of public trust and provide policymakers with more timely, actionable intelligence about the trust landscape within their policy areas.