AI in French politics

Nearly one in two French citizens has used or is considering using artificial intelligence to inform themselves about politics, according to a study by Ipsos BVA for the Jean-Jaurès Foundation. This trend highlights AI's growing role as an influential actor in the electoral process.

Youth and AI engagement

During the 2026 municipal elections, 16% of French citizens used AI to help choose their candidate, with the proportion rising to 33% among those aged 18-24, as found by Harris Interactive for Terra Nova. This shows a stronger reliance on AI among younger voters.

AI's upstream influence

AI's influence primarily occurs before voting, when users ask it to explain political programs or compare candidates. According to researcher Nadia Guérouaou of the CNRS, this creates a fundamental inequality between candidates who can provide AI with their speeches and those who cannot.

Data-driven bias

This disadvantage is not only due to financial resources but is embedded in the data used to train AI models. Candidates who are better covered by the press are mechanically better represented by AI fed with published content, reinforcing existing positions rather than changing the political landscape.

Shift from SEO to GEO

SEO, or search engine optimization, which used to determine online visibility, is being replaced by GEO, or Generative Engine Optimization. This shift reveals how models like ChatGPT, Claude, or Gemini select and present information, moving the power of visibility to a few closed systems.

Invisible algorithmic battle

For candidates, the challenge now is ensuring their program is understood and presented correctly when voters query AI. This shifts the communication battle to algorithmic decisions that are not easily visible or contestable, unlike traditional editorial biases.

Lack of AI regulation

French laws regulate speaking time and campaign expenses but do not guarantee balanced AI representation of candidates. This leaves electoral pluralism vulnerable to technical decisions that are beyond democratic oversight.

Sovereignty and disinformation

The reliability of information sources for AI and the selection of content by models raise sovereignty concerns, especially when AI is developed by American companies. These technical choices can affect democratic control over information.

AI and disinformation risks

A study by the Financial Times suggests that generative AI may moderate rather than polarize, unlike social networks. However, AI can still be used to create fake content quickly, as seen in disinformation campaigns targeting French politicians in 2026.

Public concern over truth

A majority of 86% of French citizens fear a climate of permanent doubt about the truth if AI continues to influence political information. The urgency now is to establish transparency rules for AI, as it may become the arbiter of political discourse.