Generative AI, a technology with limitless potential applications, has the potential to significantly impact democracy and elections. While narrow AI tools are already in use for specific tasks like voter record reconciliation, the implications of generative AI are harder to predict. This foundational technology is more comparable to the emergence of the internet itself, capable of transforming the mechanics of elections and governance, as well as challenging information ecosystems, public trust, and opinion.
Directly, generative AI can revolutionize policymaking by offering a more accurate understanding of potential policy outcomes. For instance, organizations like Climate Change AI utilize this technology to consider the impact of extreme weather events on infrastructure design. Generative AI is also used by law enforcement agencies for surveillance and predictive policing, and by lawyers and judges to assist in tasks like filing cases and issuing court rulings.
However, concerns arise about the impact of generative AI on elections. With numerous countries, including the United States and the European Union, set to hold elections in 2024, generative AI could introduce new biases and uncertainties. Although narrow AI systems currently streamline election administration, they struggle with issues such as matching Asian names and may exhibit biases against minorities. The use of AI for tasks like signature verification in mail-in ballots also presents potential biases. Additionally, generative AI could impact electoral rules and structures, influencing factors like election competitiveness, political polarization, voter turnout, and candidate incentives. In the US, up to 90% of congressional districts are considered safe for either Republicans or Democrats, lacking genuine competition.
Drawing on generative AI, various apps already aid legislators in redistricting efforts. While these advanced AI programs could promote fairness and representativeness in the political system, they could also enable more repressive partisan gerrymandering. Generative AI has the potential to facilitate other necessary structural reforms as well, such as the adoption of ranked-choice voting, open primaries, or proportional representation. However, the interaction effects and overall impact of such changes remain unclear in different political environments.
Apart from direct effects on the mechanics of democracy, generative AI’s indirect effects are of concern. The disruption of labor markets and subsequent political consequences are expected with the advancement of AI. Furthermore, the technology may reshape the information ecosystems relied upon by governments, candidates, and voters. Generative AI can assist journalists in summarizing government hearings, organizing contributions from multiple sources, and even drafting and editing articles. However, this could also lead to further job losses in the journalism field.
While social media platforms could employ AI to moderate content and combat election disinformation, generative AI may worsen the current disinformation crisis. It has the potential to produce highly personalized and persuasive content, tested and tailored for specific audiences across all media. This could flood the political landscape with cheaply produced, misleading content that exploits psychological vulnerabilities. Foreign adversaries could also benefit from generative AI, leveraging its capacity for persuasive disinformation in multiple languages.
Nonetheless, generative AI can support pro-democracy advocates in developing anti-authoritarian messaging specifically targeted at vulnerable communities. Currently, democratic voices are outnumbered and outgunned. AI-generated content still requires platforms like Facebook for distribution. Collaboration between AI labs and social media platforms is essential to develop effective mechanisms for preventing the spread of disinformation. It is crucial to recognize that foundational language-learning models in AI carry biases inherited from human history, favoring certain cultures and potentially amplifying dangerous biases during times of increasing political polarization.
In conclusion, generative AI offers significant benefits across various fields, but its impact on democracy requires careful consideration. While it has the potential to strengthen rule by the people, it can also undermine it. Safeguarding trust is vital, and AI’s influence on democracy must be approached with caution. Ultimately, maintaining citizens’ trust in their leaders and institutions is invaluable to the core functioning of democracy.
