IA generativa en campañas electorales: identificación de patrones globales

The report analyzes 215 incidents of generative AI use across 50 countries that held competitive national elections in 2024. It finds that GenAI was present in 80% of these countries, primarily used for creating synthetic content such as deepfake audio and video. A significant majority (69%) of these incidents were categorized as harmful to the electoral process, with 20% attributed to foreign actors, all of whom demonstrated harmful intent. The document emphasizes the challenge of anonymity, as 46% of incidents had unidentifiable authors, and recommends mandatory disclosure norms and improved detection methods to protect democratic integrity.

• Generative AI (GenAI) was utilized in the campaign phases of 80% of the countries that held competitive national elections in 2024. India and the United States were particularly prominent, with 30 documented incidents each.
• The primary application of GenAI in elections is the production of synthetic content, accounting for 90% of recorded incidents. This includes the creation of audio messages, images, videos, and social media posts designed to be realistic and misleading.
• Attributing the source of GenAI content is difficult, as 46% of incidents have no identifiable author. Among those that could be traced, candidates and political parties were responsible for 25%, foreign operators for 20%, and other groups for 9%.
• A majority of GenAI applications in electoral contexts are detrimental, with 69% of incidents categorized as harmful. All documented uses by foreign actors were found to have harmful intentions, such as interfering in processes or manipulating voter perception.
• While predominantly harmful, GenAI has been used constructively by 38% of national political parties and candidates. Examples include the use of chatbots in Japan to interact with voters and the use of GenAI for civic outreach and accessibility (16% of incidents).
• Specific instances of GenAI-driven electoral interference include the invalidation of presidential election results in Romania due to coordinated manipulation, a deepfake video of a candidate withdrawing in Bangladesh, and a fake audio clip of US President Joe Biden supporting local parties in Namibia.

APA

International Panel on the Information Environment (2025). IA generativa en campañas electorales: identificación de patrones globales. https://cdn.prod.website-files.com/643ecb10be528d2c1da863cb/690a0c9286759bdd611e2a47_Final%20ES%20Elections%20SFP.pdf

Chicago

International Panel on the Information Environment. IA generativa en campañas electorales: identificación de patrones globales. 2025. https://cdn.prod.website-files.com/643ecb10be528d2c1da863cb/690a0c9286759bdd611e2a47_Final%20ES%20Elections%20SFP.pdf.

Wikipedia

{{cite report |author=International Panel on the Information Environment |title=IA generativa en campañas electorales: identificación de patrones globales |date=2025 |url=https://cdn.prod.website-files.com/643ecb10be528d2c1da863cb/690a0c9286759bdd611e2a47_Final%20ES%20Elections%20SFP.pdf |access-date=23 October 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{internationalpanelontheinformationenvironment2025generativa, author = {{International Panel on the Information Environment}}, title = {{IA generativa en campañas electorales: identificación de patrones globales}}, institution = {International Panel on the Information Environment}, year = {2025}, url = {https://cdn.prod.website-files.com/643ecb10be528d2c1da863cb/690a0c9286759bdd611e2a47_Final%20ES%20Elections%20SFP.pdf}, urldate = {2026-10-23}, note = {Indexed by Climate Insights Directory} }