Online discussion platforms, digital deliberation tools, and – more recently – tools using artificial intelligence are increasingly being explored as ways to involve citizens more directly in policymaking processes. But can technology truly improve democratic participation? And what risks come with handing more of these processes over to algorithms and digital platforms?
The European Parliament’s Panel for the Future of Science and Technology (STOA) commissioned a study to analyse how digital tools and AI can be used to promote citizen participation in EU policymaking. ESMH spoke with one of the authors, Bjørn Bedsted, to discuss the findings of the study: what works, what doesn’t, and why the success of these initiatives depends far less on technology itself than on the democratic structures surrounding it.
Bjørn Bedsted will also speak at the STOA workshop ‘Civic AI in the EU: strengthening citizen participation through digital tools and AI’ on 9 June in the European Parliament.
Why is it important to engage citizens in EU policymaking?
Engaging citizens can help increase trust in institutions and the legitimacy of policy decisions. When citizens’ wishes and concerns are taken into account, there is also a higher likelihood of successful implementation.
What concrete measures could the EU take to improve its use of digital tools for civic engagement?
Bjørn Bedsted: It’s not really about the tools themselves. It’s about introducing citizen participation for more purposes and in more stages of policymaking. Digital tools alone cannot bring about the necessary change: what matters are the institutional conditions surrounding them.
What are the most common mistakes made when implementing digital tools to foster citizen participation?
Bjørn Bedsted: One common mistake is putting the solution, method, or tool before the problem. Institutions should first analyse what is needed and what kind of process they require. Only then should they choose the method and the tool.
Another common issue is underestimating the human resources required. People often think that because something is online, it must be easy, but it’s not. It requires a great deal of effort to make people participate successfully. There’s an entirely new skill set involved in designing and managing digital participation processes properly.
I think digital tools are best used in combination with face-to-face deliberation (consultations, debates etc). They are not very good as stand-alone tools.
What trade-offs need to be considered when deploying AI tools?
Bjørn Bedsted: One clear advantage of AI is multilingual deliberation: translation functions can help facilitate participation across languages. There are also opportunities in knowledge curation: making sense of large quantities of information.
There are, of course, risks involving privacy and other concerns. But I think the main problem arises when AI is used for synthesising and summarising deliberations. In that case, you are taking agency away from citizens, and that agency is precisely what citizen participation is about.
You might ask citizens to share their opinions during an online event or a large-scale survey. Everyone writes what they think, and then an AI system summarises and synthesises the responses into a final result, saying: “This is what citizens think.” But that is not what citizen participation is supposed to be. Citizens deliberate, negotiate, and collectively define the outcomes of their discussions. If AI simply synthesises the results for them, that undermines the entire purpose of participation.
The comparison I would make is with elected politicians themselves. If we accept that AI can simply synthesise citizen deliberation, then by the same logic, policymakers are no longer needed either. It is essentially the same principle.
Finally, what does the use of digital tools teach us about citizen engagement more broadly?
Bjørn Bedsted: I think the main lesson is that deliberation itself is central to democratic renewal. Digital tools that merely collect large quantities of citizen input – which are then analysed by public authorities, researchers, or AI systems – miss the essential point. The important part is the deliberation itself. That’s what we need to keep in mind when discussing AI and digital participation tools.