Author: Pravar Petkar, Head of Strengthening Democracy Desk
As we near the end of the ‘year of elections’, liberal democracy’s report card is a decidedly mixed one. Although there have been peaceful transfers of power across the world, elections in the UK and France have raised issues of legislative representation and coalition formation. The recent US election has seen misinformation and polarisation in campaigning, and India’s elections continue to be vulnerable to vote buying and identity politics. A recent roundtable discussion hosted by the ICfS considered how we might improve democracy with the aid of emerging technology in this climate. The discussion highlighted that countries across the world should experiment with new ways of doing democracy at the grassroots level, supported by technology. This innovation must nonetheless be accompanied by greater enforcement of existing regulations and more agile development of rules, given the real risks posed by the rapid growth of AI.
Participation is the way forward – at the grassroots level at least
Trust in the current model of popular government has been falling in some of the world’s most long-standing democracies, including the UK and the US. This is for several reasons. There are structural flaws in the election systems of representative democracies: the UK election saw a large majority on little over one-third of the vote, whilst discussion of the US election was dominated by votes in seven key ‘swing states’. These are perhaps exacerbated by misinformation, as seen in the US election campaign. Moreover, once representatives are elected, they must deliver on their promises: yet in the UK in particular, public service delivery has deteriorated significantly over the course of decades. These various systemic flaws require new ways of ‘doing democracy’ to restore that trust.
After a successful citizens’ assembly on abortion in Ireland led to its legalisation in 2018, and the open-source Pol.is platform was used by the vTaiwan initiative to generate popular consensus on Uber licensing, interest has grown in finding new ways to directly engage citizens in public decision-making. Yet putting an issue to the people at national level does not always produce change, as shown by the French Citizens’ Convention on the Climate, amongst others. The design of these mechanisms is key, and not all topics are suitable. Putting major issues of economic policy to a participatory assembly, for example, could have adverse effects for market confidence. Local democracy, then, is perhaps the best starting point, with several instances of citizen participation at local council level in recent years. Evidence of success at local level is needed to make a stronger case for it at the national level.
The benefits of participation are not limited to better or more inclusive decision-making. The polarisation that affects democracy in the UK, US and elsewhere can, in part, be put down to a lack of faith in our fellow citizens to make sound decisions: in the US, many Republicans and Democrats view each other as a ‘threat’. Addressing this requires a means of fostering genuine citizenship, where all those living in a society view themselves as engaged in a common enterprise. Well-designed participatory processes can help here by directing citizens towards engaging with each other on a shared problem.
Technology is here to stay
Alongside participatory experiments, there is growing innovation in ‘DelibTech’ – technological solutions that facilitate deliberation and discussion between citizens. In Libya, the UN used the Remesh AI platform as part of its peacekeeping efforts to gather and cluster opinions and create ‘bridging statements’ to generate consensus amongst society, with the proposals put to a Government of National Unity. The Consensus AI tool has also supported the Conference on the Future of Europe, offering participants the opportunity to vote on statements and submit their own until a 60% majority in favour of one has emerged. The Decidim platform was used in the French Citizens’ Convention on the Climate for external contributions. Greater experimentation with these tools at local level – beyond simply holding deliberative assemblies over video-conferencing – can create the groundswell for future change.
The cost of participatory democracy means DelibTech will have an important role in scaling up these processes in due course. For example, Demos estimates a single 8-day citizens’ assembly with 100 participants would cost between £800k and £1.2m. This high level of expenditure would mean that holding multiple in-person citizens’ assemblies of this kind each year would quickly become financially unsustainable; cheaper options are required, whilst preserving the ability of all citizens to engage in democracy without compromising on the income needed to maintain a reasonable standard of living.
Mechanisms such as Pol.is, Remesh and Consensus AI are also different to social media sites such as Facebook and X: whilst participants can ‘upvote’ others’ comments and add their own, the conversation ‘threads’ that can often generate hostility on other platforms are absent. Whilst many may hold the view that this is good enough reason to avoid technology entirely, a tech-free future is hardly realistic: indeed, the recent migration of social media users from X to newer platforms such as BlueSky only reinforces that technology is here to stay.
The future of democracy requires better tech regulation and education
Whilst emerging technology can play an important role in opening up new opportunities for citizen participation, it is not without its own risks and challenges. Since Large Language Models (LLMs) such as ChatGPT became widely available from late 2022, concerns have been raised around cybersecurity, data bias, privacy, copyright and the dissemination of false information. Questions were raised in this roundtable discussion about the ability of regulatory systems to keep up with rapid AI development, and whether existing regulations could be better enforced. There is also the risk that as systems such as Pol.is and Consensus AI grow, they will be taken over by large corporations. As Marietje Schaake has argued in The Tech Coup, companies such as Microsoft, Google and Meta can evade accountability from nation-states because of the practical power they wield and their influence within governments. Were DelibTech to go the same way, we would be handing over the keys of democracy to actors whose primary interest is in making profit – and who are not necessarily oriented towards the common good or responsive to citizens’ needs.
Any response must balance the innovation required for DelibTech to facilitate citizen participation with the need for intervention to address the challenges above. Four avenues for this stand out:
- Greater international cooperation on AI regulation and enforcement. This both acknowledges the cross-border nature of the technology and can prevent developers from ‘forum shopping’ in jurisdictions with limited regulations.
- More stringent antitrust regulation of AI. This will address concerns about AI monopolies created by open-source start-ups (such as Mistral in France) partnering with Big Tech companies such as Microsoft.
- Restrictions on AI development by for-profit entities. Open AI’s new for-profit status (it was previously a non-profit) opens the door for AI safety to be subordinated to profit-making considerations. Centring non-profits in AI development could guarantee greater safety, though may slow innovation.
- Improved AI literacy education. Courses such as those offered at the University of Helsinki are vital in equipping citizens with the awareness and skills to use AI responsibly. This can work hand-in-hand with regulatory responses.
Democracy has already evolved from ancient self-governing assemblies to modern representative legislatures. Further evolution is required in the age of AI to reinforce trust in the very idea of self-government. It is a path on which humans and technology must walk side-by-side: attention must now turn to experimentation, regulation and education to ensure democracy can be sustained in this emerging order.
