How is AI contributing to undermine democracy

Abstract

This essay examines the decline in Western democracies and how AI contributes to it amplifying its vulnerabilities, especially economic stagnation, such as the 2008 financial crisis, and the political polarization most of these countries experience as of today, exploring how deepfakes or bots are used to contribute to the phenomenon called ‘Truth Decay’ by interfering in foreign electoral processes and trying to disrupt public opinion by diverting its course toward certain topics during campaign periods.

Introduction

The aim of this essay is to explore the relationship between Artificial Intelligence (AI) and the decline of democracy in Western countries. My interest in this topic developed while I was studying the functioning of Large Language Models (LLM), such as ChatGPT or Gemini, during my exchange at the University of Waterloo.

Especially after World War II, Western countries — such as the US, the UK, Canada, Australia, New Zealand, and most of Europe— have lived under democratic systems, all of which share a set of common policies considered the foundational pillars of modern democracy, and used to define the notion of democracy nowadays: separation of powers, accountability, respect for human rights and fundamental freedoms— such as freedom of speech, opinion, association—, transparency, rule of law, independence of the media, political plurality, and the exercise of the power by the population through representatives elected in free and fair elections.

Democracy Deteriorating

Over recent decades, these countries have not only preserved their democratic systems, but have also made significant strides in improving quality of life and expanding the rights available to their citizens — including marriage equality, reproductive rights, protections for victims of domestic violence, and safeguards for ethnic minorities. However, over the last two decades — beginning with the 2008 global financial crisis — many of these countries have experienced a marked decline in at least one of these areas: quality of life, social mobility, trust in institutions, civic participation, corruption perception, social cohesion, fairness in electoral processes, press freedom, and rule of law.

Diving deeper into the economic issues, some countries have not yet recovered their income and quality of life levels before the 2008 crisis. While we see cases such as Greece, which lost 25% of its pre-2008 GDP per capita, most European countries have experienced stagnation in their GDP per capita, with low to no-growth or even loss, such as Italy, Spain, France or the United Kingdom, also boosted by low productivity growth.

These longstanding economic grievances bear significant responsibility for the erosion of democracy. During prolonged periods of economic stagnation, citizens experience a deterioration — or at best a stagnation — of their living standards, beginning to blame institutions they perceive as incapable of reversing the trend. As people feel more and more unfairness, polarization grows as societies break apart along the lines of their most important problems.

The role of Artificial Intelligence

In the meanwhile, AI has emerged as a major driver of democratic decline, as actors increasingly take advantage of its capabilities to undermine democratic institutions. Nevertheless, rather than representing an entirely new threat, AI appears to amplify the challenges liberal democracies have been facing for years.

The foreign disruption of electoral processes, a main concern for most countries, is a well-known example of how AI can be used to further interfere in those processes. While traditional interference tactics such as fake emails—such as allegedly during the 2016 US presidential election— or covert funding of political actors are well documented, the use of AI-generated content — such as deepfakes— represents a concerning risk, as their remarkably high quality makes it increasingly difficult to distinguish fact from fiction, especially when existing countermeasures remain underdeveloped.

Deepfakes are broadly defined as AI-generated audio or video files — the term is a portmanteau of ‘deep learning’ and fake—, intending to impersonate a real person.

A well-known example of this is the 2023 Slovak parliamentary election. During the campaign, the polling frontrunner candidate was affected by a high-quality deepfake in which he was depicted discussing a bribe, and ended up being the runner-up, losing to pro-Russian, Eurosceptic left-wing candidate Robert Fico, currently in office as of April 2026, and who strongly opposed providing aid to Ukraine — whether in the form of military equipment or loans — within the EU.

Additionally, the alleged use of bots in social media — with growing evidence of funding and setup by foreign actors — seeking to disrupt electoral processes, such as the 2025 Moldovan parliamentary elections, or further deepen polarization in certain countries, is another issue in which AI has the potential to further destabilize democratic processes.

As a result of high-quality deepfakes and fake news, our society and institutions are experiencing a loss of the baseline of truth. Fake news and media quality are so high they look real, as AI has been improving the quality of the content it is able to generate, causing a struggle to discern what is real and what is not. Although experienced users are equipped with the skills to discern AI-generated content from actual user-generated content, most people, with no background on those technologies, lack the knowledge to do so.

That said, the core issue lies in the fact that such trends have taken root in societies where misinformation — with fake news as a prime example — was already widespread long before LLMs became mainstream. These technologies can be easily weaponized to amplify misinformation, posing a fundamental threat to the core pillars of modern democracy, particularly transparency and accountability.

Where does this lead to: Truth Decay

This longstanding rampant misinformation is one of the symptoms of a phenomenon called ‘Truth Decay’, a term coined by Jennifer Kavanagh and Michael D. Rich, co-authors of the namesake book, published by RAND. It is partly defined by an increasing disagreement about objective facts, driven by cognitive biases, political and social polarization, demands on the education system slowing its ability to evolve and adapt to changes in society, as well as changes to the information ecosystem as the ones our societies have experienced and still are experiencing.

While mainly used to talk about the particular case of the United States, the term is also suitable for reference to other countries, such as Spain, where similar dynamics have been taking place, reshaping public discourse and influencing the behaviour of political actors.

Conclusion

In summary, AI is a disruptive force reshaping politics across entire societies. While its benefits are clear, so are its risks when not deployed in well-intentioned ways. Governments and civil society must therefore strike a careful balance, establishing regulations robust enough to prevent harm, yet measured enough to preserve innovation. Overly restrictive frameworks risk driving key actors away, ultimately denying society the very benefits AI has to offer.