In a new report, Dr Amélie Hennemann-Heldt, TUM Think Tank Fellow of Practice, sheds light on the nuanced role of generative AI (genAI) in elections. With over 70 national elections held in 2024, covering about half the global population, concerns around genAI’s potential to disrupt democratic integrity were at an all-time high –particularly in Germany. However, this report challenges the assumption that genAI’s impact is solely negative, revealing a more complex landscape where risks coexist with opportunities for democratic innovation.
Beyond the Fear Factor
Real-world election experiences helped to temper fears that genAI would overwhelmingly disrupt electoral processes. While some malicious actors did use genAI to manipulate information, the report found that these cases were typically isolated incidents rather than coordinated campaigns. Most misleading AI-generated content appeared to stem from careless use rather than systematic manipulation.
A tangible impact we may observe is genAI’s potential to amplify and accelerate the spread of political information. The efficiency, affordability, and scale of GenAI has transformed social media advertising and data-driven campaigning techniques. While genAI can make disinformation campaigns more effective by automating content production for multiple audiences and through customized emotionally engaging content, it also offers benefits like improved access to political information for citizens and lowered campaign costs for smaller political parties.
Regulatory Gaps and Challenges
The report also highlights shortcomings in the existing regulatory framework. Current regulations primarily address potential risks, such as misleading AI-generated content or robocalls, rather than proactively guiding the ethical and effective use of genAI in elections. The effectiveness of existing laws remains uncertain as specific genAI regulations are either nonexistent or still in early implementation phases. While major AI companies have taken steps toward self-regulation, the long-term success of these measures depends on their continued commitment. Recent trends in the US indicate a decline in transparency, making it harder for researchers to assess genAI’s influence.
Room for Innovation and Safeguards
As discussions on genAI’s impact on elections continue, democracies can benefit from balancing its potential for voter engagement with strong safeguards. Innovative civic tech solutions can help restore trust in the information ecosystem and electoral processes.
Industry collaboration and clear standards can be key to addressing immediate challenges while moderation models can aid in curbing AI-driven deception while preserving (online) space for political discourse. In terms of social media platforms, aligning policies on genAI in political campaigns and implementing rapid-response mechanisms for AI-related incidents alongside existing measures can boost trust and safety.
The report is the result of a project conducted by Dr Hennemann-Heldt at the TUM Think Tank. It is based on extensive discussions with leading experts in the second half of 2024 and identifies key trends and shifts in the discourse surrounding genAI. While early 2024 was dominated by fears of widespread AI-driven electoral interference, a shift in perspective is noted following the largely uneventful European Parliament elections. As a result, conversations moved beyond risk-focused narratives to a more balanced assessment of genAI’s applications. With Germany’s upcoming national elections, this report highlights the importance of ongoing discussions about technology’s impact on democracy.
The full report is available for download here.
Cover image by Element5 Digital via Unsplash.com