Democracy in the Age of Artificial Intelligence

We are at a turning point as a society. In recent years, advances in science and technology have redefined what we thought was possible: robots that will replace humans in manual and intellectual tasks, colonization projects on the Moon and Mars, artificial intelligence factories orbiting the Earth as satellites, direct use of solar energy in space, and unprecedented progress in medicine and biogenetics.

But the topic I want to address today—in this first installment of a series—is how Artificial intelligence is already transforming political processes, specifically in one of the most important pillars of democracy: elections.

A tool, not an end.

Before continuing, a clarification is necessary. Those familiar with my background know that I am a disseminator of knowledge and technologies, and that I consider it essential for this knowledge to reach as many people as possible, especially young people. However, I also want to be emphatic: Technology is not an end in itself, but a means to improve people's lives.

Like any tool, artificial intelligence can be used constructively or destructively. Maintaining a clear ethical framework for its application is essential, and that will be a central theme in a future installment. For now, the aim is to share three specific cases that already demonstrate the reach of AI in the electoral arena and allow us to reflect on what lies ahead.

Case 1: South Korea – Presidential Elections 2022

The candidate Yoon Suk-yeol set a global precedent by launching “AI Yoon”, A deepfake avatar created by the company DeepBrain AI. This was generated from 20 hours of audio and video recordings of the candidate himself, with about 3,000 spoken sentences.
For two months leading up to the election, the campaign team selected one voter question each day and drafted a response for the avatar to present. Initially, the content focused on public policy. Over time, the tone became more informal: “AI Yoon” shared his Myers-Briggs personality type, his favorite karaoke songs, and occasionally made sarcastic remarks about the opposing candidate.

The result was extraordinary: millions of views and a direct connection with young voters, a key demographic for his victory. Importantly, there was no intention to deceive; viewers knew they were interacting with a computer-generated animation. The tactic was so effective that his opponent, Lee Jae-myung, ended up creating his own avatar two weeks before the election.

Case 2: Pakistan – General Elections 2024

The case of Imran Khan It is one of the most emblematic examples of how AI can alter the rules of the political game. The former Prime Minister of Pakistan was imprisoned, and his party, the PTI (Pakistan Tehreek-e Insaf), was virtually silenced by the mainstream media. Faced with this, he turned to generative artificial intelligence to campaign from prison.

Khan used a clone of AI-generated voice to create campaign speeches. He started with texts he wrote from his cell, which were then reviewed and approved by his lawyers. His party organized virtual rallies broadcast on social media, since traditional media outlets were prohibited from using his name or image.

The most shocking moment came on February 9, 2024. That day, his official YouTube channel posted an AI-generated “victory speech.” In it, Khan congratulated his supporters on the massive voter turnout and denounced alleged manipulation of the results. The video was clearly identified by his party as the “AI version.”.

This case presents an interesting dilemma. Deepfakes are often associated with disinformation, but here AI became a tool for an imprisoned political leader to exercise his right to communicate with the electorate.

Case 3: Mexico – Elections 2024

In the context of the largest elections in Mexico's history, the National Electoral Institute (INE) launched an innovative institutional approach: “"Agnes"”, an AI-powered chatbot integrated into WhatsApp.

Inés was created in 2021, but was renewed for 2024 as part of the initiative “INE Certainty 2024”,” a multidisciplinary project against electoral disinformation. In this new version, the chatbot guided citizens on how to vote and where to find their polling place. It also allowed users to send suspicious texts, audio, videos, and images for real-time verification.

The initiative was made possible through a strategic alliance between the INE (National Electoral Institute), Meta (WhatsApp's parent company), and fact-checking organizations such as AFP. According to Iván Flores Ramírez, spokesperson for the INE, "this strategic alliance reflects our firm commitment to certainty and transparency in these elections.".

To use Inés, all you had to do was send the word “"Hello"” To contact the INE's WhatsApp number, select the "Certeza INE 2024" option from the menu. This case demonstrates that artificial intelligence in elections is not only in the hands of candidates and parties, but also of the institutions that protect the integrity of the democratic process.

What do these three cases tell us?

These three examples illustrate the versatility of artificial intelligence in the electoral arena. Its use is not monolithic: AI can serve as a campaign tool, a vehicle for citizen participation, or an institutional mechanism to protect democracy.

A global landscape in transformation

The impact of AI on elections is not limited to these three cases. According to an analysis by Surfshark, 38 countries have faced deepfake incidents linked to electoral processes. Collectively, they affect a population of approximately 3.8 billion people.

In 2024, over 801% of the countries that held elections reported some form of AI use in their processes. In 901% of those cases, it was used to create content.

Social media is the main channel of dissemination. They are present in the 92% of the election deepfake incidents documented. Messaging applications such as WhatsApp and Telegram (32%) and targeted online advertising (32%) are also involved.

He Brennan Center for Justice It warns that, without decisive action, “AI-driven deceptive practices could become a constant feature in political campaigns and erode the very foundations of democratic governance.”.
A particularly serious case occurred in Romania. The Constitutional Court annulled the 2024 presidential elections after discovering interference based on advanced digital technologies, including artificial intelligence.

This conversation has only just begun.

Artificial intelligence will continue to transform the way politics is conducted, but the question is not whether it will continue to impact democracy, but rather how societies will respond. Regulation, digital literacy, and transparency will be fundamental pillars for navigating this new reality.

Future installments will address the ethical framework of AI applied to politics and other dimensions of this technological transformation that is already irreversible.

---

Do you know of other cases where AI has impacted electoral processes? Share your experience in the comments, spread this article, and contribute to a conversation that involves us all.

---

Sources consulted

• How South Korean President's Campaign Used Deepfakes – DeepLearning.AI​

• The Effect of AI on Elections Around the World – Brennan Center for Justice​

• Inés, the chatbot that verifies information about the elections in Mexico – TRT Spanish​

• Imran Khan: Pakistan ex-PM used AI to campaign from jail – BBC News​

• 38 countries have faced deepfakes in elections – Surfshark​

• Gauging the AI Threat to Free and Fair Elections – Brennan Center