In the pivotal AI election year of 2024, understanding how Artificial Intelligence (AI) can change the character of election campaigns is more essential than ever. This subject is at the heart of this Young DGAP Fellowship project, which focuses on the quantitative analysis of AI’s impact on political campaigns and the resulting shifts in citizens’ trust in government and media. Within the scope of this project, Louisa Brög has developed a methodology to quantitatively measure and compare the influence of AI on elections across different countries.
Share
To date, the literature lacks a large-N quantitative evaluation of the impact of generative AI on elections. This project aims to make a first step in this direction, outlining a framework for how to measure the effect of generative AI on elections in the context of an extraordinarily large number of elections in 2024 using proxy variables. The primary focus will be the use of generative AI to create disinformation. While disinformation during electoral campaigns is nothing new, generative AI may significantly decrease the needed resources for such operations, increasing the potential impact of hostile actors on elections in 2024.
Case Study: News Coverage of AI Usage in Slovak and Polish Elections 2023?
The following graph shows news coverage about artificial intelligence in the context of the 2023 Slovak Parliamentary elections, which took place on 30 September 2023. News articles are recorded in the six months preceding and 72 hours following the elections. Hover over the peaks to find out more about this coverage.
It’s important to note that Slovakia has a 48-hour moratorium ahead of the election, during which the media and politicians are not supposed to discuss the election. This meant that the spread of a faked audio recording featuring Michal Šimečka was not adequately reflected in the data set.
The following graph shows news coverage about artificial intelligence in the context of the 2023 Polish Parliamentary elections, which took place on 15 October 2023. News articles are recorded in the six months preceding and 72 hours following the elections. Hover over the peaks to find out more about this coverage.
How did trust in government and media develop in the same period?
The following graphs show the percentage of respondents who indicated that they tend to trust the media and the government in the Eurobarometer Surveys 98, 99 and 100. These surveys were conducted in January/February, May/June and October/November respectively.
It’s important to note that the changes in government in both Slovakia and Poland might have impacted the levels of trust in both institutions for the survey conducted after the two elections in October/November. For future case studies, elections where there is no change in government might provide a better indication of the effect of coverage of the use of AI in election campaigns on citizens’ trust in media and government.
Related content
