Overview

The capacity of citizens to exchange information with government is an essential condition for democratic governance. Citizens need information to be aware of benefits and risks, maximise their well-being and safety, see what is done on their behalf, and punish governments retrospectively. When citizens exercise their voice and participate directly, it can help governments to monitor and understand social signals and societal trends, govern intelligently and efficiently, and provide collective goods.

Widely used and increasingly AI-powered digital platforms, such as social media and large language models (LLMs), provide citizens with greater capacity to both receive and provide information than ever before. Citizens can search the world’s information banks, spread information themselves, make their voices and opinions heard on an ongoing basis (not just at election time), undertake ‘tiny acts’ of democratic engagement and even collectively force policy change.

However, democracies are also at risk of widespread pollution of the information environment, with the increasingly wide use of AI by platforms and across society, through the release of consumer-facing LLMs. Over 90% of citizens report seeing misinformation and actors in information networks and are concerned or ‘very concerned;’ about deep fakes, although experience of damaging deep fakes is relatively low (15%). Such an environment could lead us to a state where people distrust all information, including official information from government itself.  In such an environment, it is crucially important to understand where citizens do find and trust government information, and when they feel emboldened to speak to government, or to complain or raise concerns when things go wrong.

Overall, we know empirically rather little about changing patterns of information exchange by citizens. This project aims to identify and improve ‘what we do know’ about trends and patterns in democratic engagement. It will do this by asking two key research questions:

• RQ1: What are the emergent trends and patterns of democratic information exchange between citizens and government, and how do they relate to changing levels of societal AI use?
• RQ2: What might incentivize citizens to interact more? Can we identify ‘nudges’ that we might apply to citizens that encourages them to interact with government, or ways of presenting information so that citizens trust it as an official source? Can we design trustworthy AI-powered interfaces for multi-turn ‘conversational consultations’?