The deepfake elections are here

Sheikh Hasina, the newly elected Prime Minister of Bangladesh and Chairperson of Bangladesh Awami League, gestures during a meeting with foreign observers and journalists at the Prime Minister's residence in Dhaka, Bangladesh, January 8, 2024. REUTERS/Mohammad Ponir Hossain

Sheikh Hasina, the newly elected Prime Minister of Bangladesh and Chairperson of Bangladesh Awami League, gestures during a meeting with foreign observers and journalists at the Prime Minister's residence in Dhaka, Bangladesh, January 8, 2024. REUTERS/Mohammad Ponir Hossain

Deepfakes can influence voters and affect election results, and diminish trust in information on which democracies thrive

Nighat Dad is the founder Digital Rights Pakistan

More than 50 countries hold elections this year, determining the future of half the planet's population. While there are debates around policy and the candidates from Indonesia to the United States, the one issue that arguably looms above everything else is the use of existing and emerging technologies such as artificial intelligence by political parties and their supporters.

The new Global Risks Report by the World Economic Forum warns that misinformation and disinformation are the most significant short-term risks. The spread of misinformation and disinformation, driven in part by new AI tools, is a key risk to major elections in 2024 that could potentially undermine the legitimacy of new governments, it said.

Political and technology experts have been having high-level discussions on how the easy access to sophisticated AI and generative AI tools will play out across the globe, and we have already started seeing examples of these technologies being employed by political parties in the hopes of influencing voters.

A recent investigation by the AFP news agency found that numerous articles lauding the policies of the Bangladeshi government led by Sheikh Hasina, purportedly written by independent experts, were published in both national and international media. However, these authors possess dubious credentials, use counterfeit photographs, and might not even be real individuals.

Hasina swept to a fourth straight term in Sunday's general election that was boycotted by the main opposition and drew a low turnout.

The websites of the three largest Political Action Committee (PAC) groups supporting U.S. President Donald Trump and his agenda, are seen on computer screens in this photo illustration taken in Washington, U.S., June 27, 2017
Go DeeperAre AI deepfakes a threat to elections?
Agustin Romo, 27, a legislative candidate for Buenos Aires province, who informally coordinates the social media team of Javier Milei's La Libertad Avanza party, works at the party's office, in Buenos Aires, Argentina August 25, 2023. REUTERS/Agustin Marcarian
Go DeeperAre social media platforms ready for record elections in 2024?
Women wait in a queue to cast their votes at a polling station during Rajasthan state assembly election in Ajmer, India, November 25, 2023
Go DeeperDeepfakes deceive voters from India to Indonesia before elections

One border over, in Pakistan, ahead of a general election scheduled for Feb. 8, the media team of incarcerated politician and former Prime Minister Imran Khan appears to be using generative AI to advance their campaign. AI-generated audio of Khan's voice accompanied a video clip that was broadcast during a virtual rally.

Although the audio featured a disclaimer indicating its synthetic nature, it still prompted questions about the appropriateness of Khan delivering a speech while in prison, as well as the audience's capacity to differentiate between synthetic and genuine media.

More recently, an op-ed by Khan surfaced on The Economist under his byline, which he later claimed was written by AI.

The party's media team then said it was, in fact, written by Khan.    

Developments such as these so close to polling day raise serious concerns around the use of tech and AI during elections not only in Pakistan, but globally.

As political parties, candidates and their supporters use AI copiously, there is a prospect of widespread confusion and distrust that can lead to an erosion of the public's trust in the integrity of the information they consume.

How election observers will measure the impact of the use of synthetic media, and how the narrative generated by synthetic media influences the mindset of voters who are already emotionally charged, remains to be seen.

It also raises questions about the role of mass media and renowned publications, and their editorial policies and fact-checking capabilities to determine if an opinion piece is generated by AI. Khan's piece in The Economist has been liked and shared by thousands of users on X and other social media platforms.

What is clear is that there are many questions around ethics and propriety that are as yet unanswered on the use of AI in the context of elections, especially in jurisdictions where the masses are not digitally literate, and where social media platforms such as TikTok, Facebook and YouTube make up the entirety of the internet experience and serve as the main source of news.

Political actors and others using AI in these countries have no accountability, with the lack of transparency in political parties, tech companies and mass media contributing to weakening already crumbling democracies.

The harsh truth is that deepfakes and synthetic media can indeed influence the mindset of voters, and eventually affect the results of elections, diminishing trust in reliable information on which democracies thrive.

Any views expressed in this opinion piece are those of the author and not of Context or the Thomson Reuters Foundation.


  • Disinformation and misinformation
  • Polarisation

Get our data & surveillance newsletter. Free. Every week.

By providing your email, you agree to our Privacy Policy.

Latest on Context