AI’s Limited Impact on 2024 European Elections Surprises Experts

AI’s limited impact on European elections: Recent research suggests that AI-generated falsehoods and deepfakes had minimal effect on election results in the UK, France, and the European Parliament in 2024.

• Sam Stockwell, a researcher at the Alan Turing Institute, conducted a study focusing on three elections over a four-month period from May to August 2024.
• The study identified only 16 cases of AI-enabled falsehoods or deepfakes that went viral during the UK general election and 11 cases in the EU and French elections combined.
• None of these cases appeared to definitively sway the election results.

Reasons for AI’s ineffectiveness: The study reveals several factors contributing to the limited impact of AI-generated content on election outcomes.

• Most people exposed to AI-generated disinformation already held beliefs aligned with the underlying message, such as concerns about high immigration levels.
• Those actively engaging with and amplifying deepfake messages typically had preexisting views that aligned with the content, reinforcing existing beliefs rather than influencing undecided voters.
• Traditional election interference tactics, like using bots to flood comment sections and exploiting influencers to spread falsehoods, remained more effective than AI-generated content.

Current use of AI in disinformation: While AI tools were employed in some capacity, their use was limited and not significantly more effective than traditional methods.

• Bad actors primarily used generative AI to rewrite news articles with their own spin or create additional online content for disinformation purposes.
• Felix Simon, a researcher at the Reuters Institute for Journalism, notes that AI is not providing much advantage currently, as simpler methods of creating false or misleading information continue to be prevalent.

Challenges in assessing AI’s impact: Experts caution that it’s still difficult to draw firm conclusions about AI’s influence on elections at this stage.

• Samuel Woolley, a disinformation expert at the University of Pittsburgh, points out the lack of sufficient data and the potential for less obvious, downstream impacts on civic engagement.
• Stockwell acknowledges that early evidence suggests AI-generated content could be more effective for harassing politicians and sowing confusion than changing people’s opinions on a large scale.

Emerging concerns: The research highlights potential long-term risks associated with AI-generated content in the political sphere.

• Politicians, including former UK Prime Minister Rishi Sunak, were targeted by AI deepfakes showing them promoting scams or admitting to financial corruption.
• Female candidates faced nonconsensual sexual deepfake content intended to disparage and intimidate them.
• The increasing difficulty for people to discern between authentic and AI-generated content in the election context raises concerns about the integrity of political processes.

Political exploitation of AI: Some politicians have begun to take advantage of the confusion surrounding AI-generated content.

• In the European Parliament elections in France, political candidates shared AI-generated content amplifying anti-immigration narratives without disclosing its artificial origin.
• Felix Simon warns that this covert engagement and lack of transparency by political actors may present a greater risk to the integrity of political processes than the use of AI by the general population or “bad actors.”

Looking ahead: Potential long-term consequences: While the immediate impact of AI on elections appears limited, researchers warn of potential future risks to democratic processes.

• The ongoing harassment and targeting of politicians with AI-generated content could have a chilling effect on their willingness to participate in future elections and harm their well-being.
• The increasing difficulty in distinguishing between real and AI-generated content may erode trust in political information and institutions over time.
• As AI technology continues to advance, its potential to influence elections and political discourse may grow, necessitating ongoing vigilance and research into its effects on democratic processes.