The report "Voting in the Age of Intelligence," signed by Rafael Conde, analyzes the risks of AI ahead of 29N and proposes measures to verify the origin of the contents and the information of the attendees
Thinkia, an AI-Native consulting firm, has published "Voting in the Age of Intelligence: what 29N teaches about AI in electoral processes." Its main thesis is that identifying content generated with artificial intelligence is necessary, but insufficient to strengthen electoral integrity. The analysis proposes moving toward mechanisms that allow its origin to be verified, its sources to be cross-checked, and errors or impersonations to be addressed quickly.
Signed by Rafael Conde, CEO and founder of Thinkia, the document examines the campaign for the Spanish general elections from November 16 to November 29, the closing date of the report. It also collects experiences from this fall’s elections in Brazil and the United States.
Among its recommendations, digital signing of official pieces from political parties and institutions, periodic audits of attendees’ responses, and the designation of responsible parties and response protocols in the event of incidents stand out. According to the report, these measures can be addressed with technologies that are already available.
The analysis distinguishes three areas: content production, where AI facilitates the massive creation of messages, videos, and voices; intermediation, with attendees as the route for consulting electoral information; and scrutiny, where these tools can help to cross-check programs, interviews, and promises with primary sources and transparent methods.
The document brings together international evidence about the scope of the phenomenon. In the United States, an October cut of the Wesleyan Media Project identified 220 political ads generated or modified with AI, associated with more than 137 million dollars in advertising spending. Only 27% included a notice about its use. In Brazil, the VigIA project by Agência Lupa analyzed 972 suspicious political-electoral pieces published between August 16 and September 28: 920 used AI and 554 were deepfakes. Only 40% of the AI pieces included the required notice. In Spain, 62.4% of the population believes that AI facilitates the spread of false news, according to FECYT.
The report includes a risk map by scope and election phase, along with an action agenda for institutions, parties, platforms, media and citizens. The recommendations are organized into three horizons: the municipal elections of May 2027, the implementation of European obligations for high-risk AI systems, and the European elections of 2029.
"The democratic impact of AI will depend less on what models are capable of generating than on what society is able to verify," Rafael Conde says.
The full report and its executive version are available in Spanish and English at https://thinkia.com/thoughts/whitepapers/voting-in-the-intelligence-era/
About Thinkia
Thinkia is a global AI-Native consultancy specialized in the design and implementation of business systems based on artificial intelligence. With international presence, it helps organizations integrate AI into their operating model through architecture, integration, governance and transformation oriented toward real business impact.