Authors: Spencer McKay, Chris Tenove, Nishtha Gupta, Jenina Ibañez, Netheena Mathews, Heidi Tworek
As 2024 shapes up to be the “Year of Deepfake Elections,” generative artificial intelligence, or GenAI, is being used at unprecedented levels. GenAI has been used to imitate political figures, harass political candidates, and spread incorrect information globally.
This report reveals how the increasingly popular technology threatens democratic processes like elections by enabling deception, polluting the information environment, and intensifying harassment. GenAI lowers the costs of deploying existing disinformation tactics that already plague social media platforms. To date, there is little evidence showing that the beneficial uses of GenAI in elections will outweigh these harmful ones.
While the report cautions against exaggerating GenAI threats, it emphasizes that policymakers cannot be complacent.
In the absence of a technological silver bullet to counter threats to democracy, state-led regulation and electoral management oversight are necessary. The report advocates for sector-specific recommendations, urging policymakers to develop a resilient information ecosystem with institutions that not only produce accurate information but also cultivate citizens’ trust in these systems.
Report Credits
Lead authors: Spencer McKay and Chris Tenove
Co-authors: Nishtha Gupta, Jenina Ibañez, Netheena Mathews, Heidi Tworek
Report design and illustration: Oliver McPartlin
We would like to thank Emily Laidlaw of the University of Calgary and Fenwick McKelvey of Concordia University for their thoughtful feedback on an earlier version of the report.
Additional information about the research team and acknowledgments is available in the report.
Citation: McKay, Spencer, Chris Tenove, Nishtha Gupta, Jenina Ibañez, Netheena Mathews, Heidi Tworek. (2024). Harmful Hallucinations: Generative AI and Elections. Vancouver: Centre for the Study of Democratic Institutions, University of British Columbia. DOI: https://doi.org/10.14288/1.0445035
Publication info: Copyright © 2024; Centre for the Study of Democratic Institutions; University of British Columbia. This work is licensed under CC BY-NC-ND 4.0 (creativecommons.org/licenses/by-nc-nd/4.0).