Author: noirindempsey1

What do deepfakes and democracy have in common? Both draw their power from the people | By Noirin Dempsey

noirindempsey1

As AI and deepfakes edge their way into electoral processes, it is important to recognize that voters control the influence of these tools.  Understanding voters’ psychological and behavioral responses to deepfakes is required in order to develop policy that can restore trust in democratic institutions and improve public awareness of AI’s role in electoral processes.

The chances that you’ve seen a deepfake video within the past year are pretty high – 60%, according to data from Jumio.1  Today, AI and deepfakes are all over the Internet, on topics from finance to healthcare to government and politics.  According to Łabuz and Nehring (2024), deepfakes are defined as “AI-generated or AI-manipulated audio, image, or video content that appreciably resembles existing persons, objects, places, or other entities or events and would falsely appear to a person to be authentic or truthful.”2  In recent years, deepfakes have increased in quality and availability to the average person while decreasing in cost, contributing to a surge in their proliferation.3

Deepfakes have been used in the lead-up to elections in many ways, varying in believability, timing, and impact.  Most of the time, deepfakes have next to no impact on elections, such as the 2024 U.S. presidential election, when deepfake images featuring Kamala Harris with communist imagery and Trump with the Taliban proliferated online.4 In some cases, though, deepfakes are strategically employed in ways that do influence elections.  For instance, a few days before the 2023 Slovakia parliamentary election, an audio deepfake circulated on social media depicting Progressive Slovakia party candidate Michal Šimečka talking with a journalist about a “scheme to rig the election by buying votes from the country’s marginalized Roma minority.”5  Although “police concluded the recording was not persuasive enough to warrant a criminal investigation,”6 Šimečka still ended up losing the election by 5%.

For the most part, however, unless deployed on a strategic timeline (usually close to an election), deepfakes tend to have minimal impact on elections.  Studies by the Institute of Political Science at the University of Bamberg show that even when parties use deceptive AI (such as impersonating public figures), public support for the party does not change.7  The study suggests that although “partisans disapprove of deceptive AI use in principle,” they “are unwilling to penalize their own party for it in practice.”8  Therefore,  candidate favorability within one’s own party is unaffected by the use of deepfakes, suggesting that candidates’ deceptive use of deepfakes in campaigning is unlikely to lose them votes.

That is not to say deepfakes have no real effect on politics – quite the opposite.  But the deepfake’s real power comes not from the content itself, but from voters’ reactions.  Deepfakes often elicit real psychological and behavioral responses.  For one, deepfakes encourage political polarization and extreme views.  As economist Zeynep Arzu Kelani and sociologist Darina Lepadatu explain in their article on deepfake propaganda during the 2024 U.S. presidential election, “deepfakes that depict political figures in a negative light tend to disproportionately impact members of opposing parties, resulting in intensified distrust and division” along partisan lines.9  Many deepfakes are emotionally manipulative, eliciting intense reactions to “emotionally charged content” which encourage  individuals to “share misinformation that resonates with their social identity, beliefs, or social norms.”10

Deepfakes can also diminish trust in elections and democracy overall.11 After interacting with deepfakes, people often begin to view legitimate government content with skepticism, including election results.  Fola Adeleke, Executive Director and co-founder of the Global Center on AI Governance, notes one key example that occurred in the 2022 Kenyan Supreme Court case Obinga v. Ruto.12  In this landmark case, the court ruled that the Kenya Integrated Electoral Management System, or KIEMS (a technological tool used for voter registration and identification, as well as transmitting election votes to be counted), was constitutional and had produced accurate election results for the 2022 presidential election.13  Despite the court’s ruling, however, Kenyan citizens largely distrusted the election results and questioned the government’s transparency in using KIEMS.14  Such citizen distrust can be dangerous because it often means that “political participation declines and government finds it difficult to command…legitimacy and respect,”15 thus contributing to the erosion of democratic institutions over time.

As AI and deepfakes continue to take on an ever greater role in government processes, leaders must develop policies focused on a framework centered on the human person.  Politicians must assure citizens of accountability and transparency in elections, and one way to do this is by creating policies that review election technologies to verify their ethicality.16  Adeleke also suggests that political institutions should consider partnering with private corporations such as nonprofits to run elections.  Doing so will increase transparency and trust in government, since multiple organizations will be involved in the electoral process, many of which the public see as “trustworthy.”17  Additionally, it is important for governments and citizens to recognize that people do not have equal access to technologies and may be illiterate.18  As such, investing time and resources in public education on AI and other technologies could be a helpful way to build public trust in government institutions. Additionally, education resources can promote acceptance of certain uses of AI while instituting legal consequences to dissuade people from unethical practices such as spreading misinformation through deepfakes.

Despite deepfakes’ inability to influence election outcomes on a large scale, their ability to elicit a psychological and emotional response should not be understated.  Nevertheless, the key takeaway is that deepfakes draw their power from the voters.  Therefore, it is up to the voters to decide how much power to give to them in our democratic institutions and whether we allow them to influence the way we approach and view our electoral processes.

  1. Jumio. “2024 Online Identity Study: Global Consumer Research.” Jumio, 2024. https://www.jumio.com/2024-identity-study/.
    ↩︎
  2.  Łabuz, Mateusz, and Christopher Nehring. “On the Way to Deep Fake Democracy? Deep Fakes in Election Campaigns in 2023.” European Political Science 23, no. 3 (2024): 454–73. https://doi.org/10.1057/s41304-024-00482-9. 
    ↩︎
  3. Ibid. ↩︎
  4.  Kelani, Zeynep, and Darina Lepadatu. “Generative Artificial Intelligence, Misinformation and Political Polarization: A Content Analysis of Deepfake Propaganda During the 2024 US Presidential Election.” Journal of Information Technology & Politics (2026). https://doi.org/10.1080/19331681.2026.2652897. ↩︎
  5. Labuz and Nehring. “On the Way to Deep Fake Democracy? Deepfakes in Election Campaigns in 2023.” ↩︎
  6. Kováčik, Timotej, and Veronika Frankovská. “How AI-Generated Content Influenced Parliamentary Elections in Slovakia: The Slovak Police Will Investigate the Recording for a Third Time.” CEDMO, November 25, 2024. https://cedmohub.eu/how-ai-generated-content-influenced-parliamentary-elections-in-slovakia-the-slovak-police-will-investigate-the-recording-for-a-third-time/. 
    ↩︎
  7. Jungherr, Andreas, Adrian Rauchfleisch, and Alexander Wuttke. “Artificial Intelligence in Election Campaigns: Perceptions, Penalties, and Implications.” Political Communication 43, no. 4 (2026): 585–606. https://doi.org/10.1080/10584609.2025.2611913.
    ↩︎
  8. Ibid. ↩︎
  9. Kelani and Lepadatu. “Generative Artificial Intelligence, Misinformation and Political Polarization: A Content Analysis of Deepfake Propaganda During the 2024 US Presidential Election.” ↩︎
  10. Ibid. ↩︎
  11. Jungherr, Rauchfleisch, Wuttke. “Artificial Intelligence in Election Campaigns: Perceptions, Penalties, and Implications.” ↩︎
  12. Adeleke, Fola. “Reinstating Trust in Elections in the Era of Artificial Intelligence and Emerging Technologies.” Data and Policy 7 (2025): e38. doi:10.1017/dap.2025.14
    ↩︎
  13. Ibid. ↩︎
  14. Ibid. ↩︎
  15. Diamond, Larry. “The Political Drivers of Low Trust.” Open Government Partnership. Accessed October 4, 2026. https://www.opengovpartnership.org/trust/the-political-drivers-of-low-trust/. 
    ↩︎
  16. Fola. “Reinstating Trust in Elections in the Era of Artificial Intelligence and Emerging Technologies.” ↩︎
  17. Ibid. ↩︎
  18. Ibid. ↩︎

What’s the Deal With Ranked-Choice Voting? | By Noirin Dempsey

noirindempsey1

For all the criticism surrounding U.S. electoral systems–whether it be gerrymandering, campaign spending, or election integrity–there is little discussion of a particular system which offers several potential improvements to the current election process: ranked-choice voting.

Many of us have heard the term, but what actually is ranked-choice voting?  Ranked-choice voting, or RCV, is a system in which voters rank their preferred candidates on the ballot by first choice, second choice, third choice, and so on.  If a candidate has the majority of first-choice votes, they win the election.  If no candidate receives a majority, the candidate with the fewest first-choice votes is dropped from the race, and the ballots which listed the dropped candidate as their first choice now contribute their second choices, and so on.  This process of removing the least-voted-for candidate continues until there is a candidate with a majority.  RCV is sometimes called “instant runoff” because there is no need for subsequent elections; voters have indicated their preferences in the event that their first choice or even their second or third choice does not get elected.  RCV is already used in municipal elections in some major US cities, including San Francisco and New York City.1

Advocates for RCV list several benefits.  For one, RCV seems to improve voter experience.  Based on a Y2 Analytics poll of Utah voters in the 2021 municipal elections, 86% were satisfied with the voting experience, reporting it was “easy to use” with clear instructions.”2 In addition, since a ranked-choice system levels the playing field for candidates, voters feel freer to choose candidates with whom they feel aligned on values rather than simply picking the person they think is most likely to win.  According to a 2022 survey conducted by Alan James Simmons et al., “Researchers asked survey respondents to vote in a hypothetical version of the 2020 US presidential election between Joe Biden (Democrat), Donald Trump (Republican), Howie Hawkins (Green), and Jo Jorgensen (Libertarian). Among respondents randomly assigned to vote under plurality rules, 3.75% voted for Hawkins or Jorgensen; among respondents randomly assigned to vote under RCV, 7% ranked Hawkins or Jorgensen first.”3  This indicates that voters are more likely to vote for third-party candidates in an RCV system.

Another benefit of RCV is that policy is more likely to reflect voter preferences.  In a “first-past-the-post” voting system like we use in the U.S., most candidates seek the support of interest groups and major political parties to propel them towards election.  As a result, such groups can motivate public elected officials to make policy decisions based on seeking party support instead of passing laws that voters want.4 RCV, on the other hand, creates incentives for candidates, once elected, to adopt moderate policy positions “with the knowledge that the extremists will be eliminated in the first round of the runoff.”5

Lastly, RCV systems seem to increase voter turnout.  According to a 2024 study of voter data conducted by Eveline Dowling and other researchers, “results show that people in RCV jurisdictions are significantly more likely to vote in off-year elections than individuals living in non-RCV places.”6  This may be because voters tend to interact more with candidates and campaigns in RCV elections.7

Not only does RCV improve the voting process, but it also seems to contribute towards a better overall experience for candidates.  In a RCV system, candidates must appeal to all groups of voters in order to win, not just their own base.  As Dowling and Tolbert explain, “this need for broader appeal can lead to more civil and cooperative campaigning.”8  Campaigns also become less negative, since candidates don’t want to “engage in negative campaigning” or malign the images of their opponents, as these actions “could alienate potential supporters.”9

RCV also encourages more candidates to run for office, as voters are free to cast their vote more sincerely.10  Unlike in a plurality or majority electoral system, candidates have a chance of winning even without the support of a major political party due to the ranking system.

Despite its benefits, RCV certainly comes with some drawbacks.  In a climate where Americans already distrust elections, decreased voter trust in the electoral system is a legitimate concern.  Derek Monson et al. found that “39% of Americans do not trust the fairness of elections in the U.S.,” which does not bode well for changing the voting system.”11  Additionally, there is evidence of voter mistrust in RCV specifically.  In a recent study, participants were given a scenario in which RCV and first-past-the-post systems yielded different election outcomes.  The study found that US survey respondents were generally more accepting of the first-past-the-post outcome.12  Implementing RCV could exacerbate voter skepticism surrounding elections, which could be seriously damaging to democracy in the U.S.

Another frequently cited concern is a phenomenon known as ballot exhaustion.  Ballot exhaustion occurs when all of a voter’s chosen candidates on a ballot have been eliminated due to runoff and there are more than two candidates left in the race.  Then, the ballot would be considered exhausted13 because it no longer contributes a tally for the final winner.

Ballot exhaustion has raised questions about whether or not RCV is constitutional.  In 2011, Dudum v. Arntz challenged San Francisco’s ranked-choice voting system (known as Instant-Runoff Voting, or IRV) for city officials.14  San Francisco law limited the number of rankings on each individual ballot to three, so plaintiffs argued that “when more than four candidates run for a particular office, the IRV excludes exhausted voters from full participation.”15  Although the Ninth Circuit Court ultimately upheld IRV as constitutional, questions certainly remain regarding whether or not RCV conforms to the Constitution in all its various forms.

Overall, RCV provides key improvements to the current U.S. voting system; however, it still has several drawbacks that cannot be ignored.  Before considering implementing RCV in major elections, more data is needed to help predict the effects such a system would have in the U.S.  Other countries such as Italy and Brazil have implemented RCV more broadly, but their government systems are much different than that of the U.S., which makes it difficult to directly compare and predict the potential consequences of widespread institutionalization of RCV in the U.S.  The public must also be educated about RCV so that they can develop well-formed opinions on the system and perhaps build more trust in its legitimacy.  Still, even if the data advocates for the U.S. to implement RCV, it is ultimately the voters’ decision whether or not to do so.  True to RCV’s principles of representation, moderation, and amplifying voters’ unique preferences, the American public will have to decide whether there is more merit or fault in RCV.

  1. Eggers, Andrew, and Laurent Bouton. “Ranked-Choice Voting.” John Hopkins University School of Public Policy and The University of Chicago Center for Effective Government, 2024. https://sgp.jhu.edu/primers/ranked-choice-voting/.
    ↩︎
  2. Monson, Derek, Josh McCrain, Zachary Taylor, and Nadia Mahallati. “The Benefits and Drawbacks of Ranked-Choice Voting in Utah.” Sutherland Institute, 2022. https://sutherlandinstitute.org/wp-content/uploads/2022/04/The-Benefits-and-Drawbacks-of-Ranked-Choice-Voting-in-Utah-2022.pdf.
    ↩︎
  3. Eggers, Andrew, and Laurent Bouton. “Ranked-Choice Voting.” John Hopkins University School of Public Policy and The University of Chicago Center for Effective Government, 2024. https://sgp.jhu.edu/primers/ranked-choice-voting/.
    ↩︎
  4. Monson, Derek, Josh McCrain, Zachary Taylor, and Nadia Mahallati. “The Benefits and Drawbacks of Ranked-Choice Voting in Utah.” Sutherland Institute, 2022. https://sutherlandinstitute.org/wp-content/uploads/2022/04/The-Benefits-and-Drawbacks-of-Ranked-Choice-Voting-in-Utah-2022.pdf.
    ↩︎
  5. Eggers, Andrew, and Laurent Bouton. “Ranked-Choice Voting.” John Hopkins University School of Public Policy and The University of Chicago Center for Effective Government, 2024. https://sgp.jhu.edu/primers/ranked-choice-voting/.
    ↩︎
  6. Dowling, Eveline and Caroline Tolbert. “What We Know About Ranked Choice Voting, Updated for 2025.” ABA Task Force for American Democracy, March 6, 2025. https://www.americanbar.org/groups/public_interest/election_law/american-democracy/our-work/what-we-know-about-ranked-choice-voting-2025/.
    ↩︎
  7. Dowling, Eveline and Caroline Tolbert. “What We Know About Ranked Choice Voting, Updated for 2025.” ABA Task Force for American Democracy, March 6, 2025. https://www.americanbar.org/groups/public_interest/election_law/american-democracy/our-work/what-we-know-about-ranked-choice-voting-2025/.
    ↩︎
  8. Dowling, Eveline and Caroline Tolbert. “What We Know About Ranked Choice Voting, Updated for 2025.” ABA Task Force for American Democracy, March 6, 2025. https://www.americanbar.org/groups/public_interest/election_law/american-democracy/our-work/what-we-know-about-ranked-choice-voting-2025/.
    ↩︎
  9. Dowling, Eveline and Caroline Tolbert. “What We Know About Ranked Choice Voting, Updated for 2025.” ABA Task Force for American Democracy, March 6, 2025. https://www.americanbar.org/groups/public_interest/election_law/american-democracy/our-work/what-we-know-about-ranked-choice-voting-2025/.
    ↩︎
  10. Eggers, Andrew, and Laurent Bouton. “Ranked-Choice Voting.” John Hopkins University School of Public Policy and The University of Chicago Center for Effective Government, 2024. https://sgp.jhu.edu/primers/ranked-choice-voting/.
    ↩︎
  11. Monson, Derek, Josh McCrain, Zachary Taylor, and Nadia Mahallati. “The Benefits and Drawbacks of Ranked-Choice Voting in Utah.” Sutherland Institute, 2022. https://sutherlandinstitute.org/wp-content/uploads/2022/04/The-Benefits-and-Drawbacks-of-Ranked-Choice-Voting-in-Utah-2022.pdf.
    ↩︎
  12. Eggers, Andrew, and Laurent Bouton. “Ranked-Choice Voting.” John Hopkins University School of Public Policy and The University of Chicago Center for Effective Government, 2024. https://sgp.jhu.edu/primers/ranked-choice-voting/.
    ↩︎
  13. Monson, Derek, Josh McCrain, Zachary Taylor, and Nadia Mahallati. “The Benefits and Drawbacks of Ranked-Choice Voting in Utah.” Sutherland Institute, 2022. https://sutherlandinstitute.org/wp-content/uploads/2022/04/The-Benefits-and-Drawbacks-of-Ranked-Choice-Voting-in-Utah-2022.pdf.
    ↩︎
  14. Balser, Jimmy. “Ranked-Choice Voting: Legal Challenges and Considerations for Congress.” Congressional Research Service, 2022. https://www.congress.gov/crs_external_products/LSB/PDF/LSB10837/LSB10837.1.pdf.
    ↩︎
  15. Balser, Jimmy. “Ranked-Choice Voting: Legal Challenges and Considerations for Congress.” Congressional Research Service, 2022. https://www.congress.gov/crs_external_products/LSB/PDF/LSB10837/LSB10837.1.pdf.
    ↩︎