In 2024, elections around the world were notably impacted by the integration of generative artificial intelligence (AI), leading to significant changes in how various stakeholders conducted their work. A new report titled "Adaptation and Innovation: The Civic Space Response to AI-Infused Elections" delves into these developments through three detailed case studies.
The year was marked by diverse uses of AI in electoral processes, including independent news delivery via AI avatars in Venezuela, the option for UK voters to elect an AI avatar to Parliament, and deepfake videos involving prominent political figures like U.S. President Joe Biden. These examples underscore concerns that AI could undermine human rights and democracy.
In response, governments from Brazil to the Philippines enacted laws regulating AI's role in elections. The European Commission issued guidelines for companies on labeling AI-generated content, while U.S. election administrators adapted communication strategies to address this new landscape.
Political campaigns globally embraced AI tools for advertisement creation and voter outreach. In Indonesia, candidates utilized ChatGPT for speech writing and campaign strategy, while in India, deepfake technology facilitated personalized voter engagement. Germany's far-right AfD party ran anti-immigrant ads featuring AI-altered images.
Despite cuts to trust and safety teams, some tech firms initiated election integrity programs. Twenty-seven companies signed the AI Elections Accord to tackle deceptive content through enhanced detection efforts. Google limited its Gemini chatbot's election-related responses and collaborated with Jigsaw on media literacy initiatives ahead of EU elections.
Civic space actors also adjusted their strategies in anticipation of increased threats from AI tools. The report examines contributions from fact-checking collectives in Mexico, decentralized civil society efforts in Taiwan, and global tracking projects by media organizations regarding generative AI use during elections.
These case studies reveal varied approaches to fostering resilient information environments amidst challenges such as funding constraints, access to information, detection difficulties, and politicization of media resilience efforts.
The findings highlight that addressing AI's role in elections is an ongoing endeavor requiring collaboration between civic groups and private entities beyond 2024. As the global political climate grows more hostile toward election-related research, continued support for information resilience is essential.
