The Impact of AI on Democracy

In the lead-up to the European elections in early June, it was hard to go a week without hearing claims that Artificial Intelligence (AI) was damaging the democratic process and feeding electorates with disinformation. Has AI really impacted our elections, or is its impact merely highlighting issues we have been trying to overlook?

We have come a long way since Barack Obama’s 2008 presidential campaign successfully used social media to engage younger voters. In 2024, technology has advanced to the point that AI can create deceptively realistic false content, such as "deepfakes" which can mislead the public about candidates’ stances on issues and events. Alternatively, AI could be used to spread disinformation to suppress voter turnout through false messages about voting locations or times or by discouraging voters from going to the polls. On the surface, it may seem that such content is aimed at infringing on voters’ right to make informed decisions. When we take a broader perspective, however, we can see that the problems people are concerned about regarding AI are highlighting a much bigger problem: technology is challenging the concept of the nation state and how individuals find an identity.

Affinities across borders
Traditionally, people would find their identity within their nation, but now they find common ground internationally. Voters in US elections are seen wearing t-shirts stating that they “would prefer to be a Russian than a Democrat.” Where the far right was often previously limited to one country, they are now finding common ground not just within national borders but beyond them in a global, internationally organized community. Digital technologies are enabling these connections and the sense that they are part of a worldwide movement – that instead of needing to find common ground with those with whom they share a geographical boundary, they instead connect digitally with others who share their values and ideas of society.

So, where Obama’s campaign was about creating a movement within a nation’s borders, today’s technology campaigns are often driven by foreign influence. Russia, China, and Iran are making well-documented and highly skilled use of social media against the rest of the world to destabilize democracy and challenge the traditional power structures of the global world order. 

At a time when digital technologies are contributing to such profound divisions in the world, it is crucial to remember that the Internet was born out of warfare – and it was as much about behavioural science as it was about protocols. The common narrative that we hear today – that the Internet was developed for openness and open societies – is, at best, disingenuous. Its very design is deeply rooted in controlling people and executing warfare on other nations. The origins of the US Defense Advanced Research Projects Agency (DARPA) project, conceived in the Eisenhower era of the 1950s, which designed and implemented the foundations of Internet technology, were developed in part with the idea of being able to control people with the new technology of computers. DARPA hoped to do this on the Russian population or, at the very least, find a way to influence those in charge of Russian nuclear weapons. The fact that Russia, China, and Iran are now finding innovative ways to deliver on this original mission should not surprise anyone.

Introducing the Datalists
While it is tempting to blame this entirely on foreign influences, the issue is somewhat more complex because without the support and infrastructure provided by several companies it would not have been possible for this foreign influence to occur in the first place. Most of these companies are based in the US. Here, we need to look more closely at the social media and high-tech companies that I call the “Datalists” – companies such as Meta, Google, Amazon, and, lately, OpenAI. These companies are fundamentally different from other technology companies in one crucial way: since the early 2000s, these companies have made a concerted effort to gather and collect data about human beings to study – and control – patterns of human behaviour. These issues are usually discussed through the lens of “privacy” invasion or targeted adverts, but it is worthwhile to think about them from a slightly different viewpoint when it comes to elections. These technologies are designed to disrupt the creation of collective memory – or the ability of people to find common ground and common, agreed-upon interpretations of what is happening in a nation and what is happening in the world. They have not been designed this way by chance. They have been designed to cause ruptures in how society organizes itself.

Some are anti-democrats
If you look into the origins of many social media and Datalist companies, there is a key theme. Several of the original funders are not necessarily pro-democracy themselves. The most-cited author among these Silicon Valley elites is Ayn Rand. Many of the founders – and crucially – the funders in Silicon Valley still religiously form themselves in the image of John Galt from Atlas Shrugged. While this style of ideology may be relatively harmless in other peoples’ hands, coupled with an extremely powerful technology that is now applied not just at a local level but on a global level to manipulate large populations of people, it raises the question of whether these people are the ones we want running the world. With the amount of data they hold on people, they can change opinions and ideas – and sway political sentiment. In short, while the Russians may be making great use of digital technologies and AI to cause rifts in democratic processes, they may just be copying the US-American Datalists’ own game.

It is also crucial to note that democracies themselves contribute to this confusion. Any time a regulator has attempted to suggest that social media should be held accountable for what is said on a platform, the Datalists avoid serious new regulations. While the EU AI Act or the Digital Services Act (DSA) attempt to change the game, governments must recognize that the concentration of power over the data that forms the basis of these technologies is critical to manage. Whereas previous monopolies attempted to control markets for products and services or the outputs in the economy, the new companies are creating monopoly control over data – a crucial input to the digital economy – and a monopoly over the ability to control human behaviour.

And then there is quantum computing
It is also important to note that AI is just one part of the spectrum of digital technologies that impact our society, economy, and environment. Many others exist: blockchain, the Internet of Things (IoT), and quantum computing. As quantum computing increasingly gains pace, the impact of AI on our democracy will look quaint. Quantum computing will increase the complexity of challenges facing our geopolitical and democratic order.

The common foundation of all these technologies is data. Rather than set regulations for every new technology that emerges, such as AI, blockchain, cryptocurrency, and presumably quantum computing, regulators and citizens must instead focus on the rules surrounding data that form the basis of our newly emerging digital economy. This requires a more detailed investigation than just the right to “privacy” or “right to be forgotten.” The rules of the game around ownership, monopoly power, and control must be addressed within the context of commonly agreed social norms. While every nation will have a tailored response to this new imperative, the conversation must begin with the question, “Who do we want to make the decisions about our country? Is it a handful of companies housed in Silicon Valley?”

Therefore, we should consider AI as illustrating and exacerbating problems rather than causing them; as always, technology highlights human fallibility. The issues we are facing with AI within our election cycles are the logical conclusions of the original DARPA project — we can no longer ignore them and hope they will solve themselves. We must execute our human intelligence to do so.