The 2026 midterm elections are rapidly approaching, and with them, a new and potentially disruptive force in political campaigning: artificial intelligence. While concerns about AI-generated deepfakes have dominated early discussions, a growing body of research suggests the more immediate threat lies in AI’s ability to actively persuade voters, shifting political attitudes at a scale previously unseen.
According to a recent Ipsos BVA survey conducted for Le Monde and the Fondation Jean-Jaurès, , more than two-thirds of French citizens (68%) have already used a generative AI tool, a figure that has tripled in the past two years. This widespread adoption is extending into the political sphere, with nearly half of French voters (48%) having either used or intending to use AI to gather information about candidates or parties.
The shift is particularly pronounced among younger voters. Ipsos BVA data reveals that 75% of 18-24 year olds have engaged with AI for political information, and 64% plan to do so during the upcoming municipal elections. This demographic, many of whom will be first-time voters, is increasingly turning to AI tools like ChatGPT (used by 50% of French citizens) and Gemini (42%) for political insights.
The ways in which voters are utilizing these tools are revealing. Approximately 39% are seeking information on key policy positions, while 31% are looking for a candidate’s overall stance on issues. Notably, 30% are considering using AI to identify the candidate most aligned with their own beliefs, and 22% are exploring AI’s potential to compare candidates side-by-side. This suggests a growing reliance on AI as a personalized political compass.
This trend isn’t limited to information gathering. The research indicates that AI is becoming a factor in shaping voter preferences. A quarter of French citizens report having already questioned an AI model about a political figure or party, and a similar proportion (27%) intend to do so during the municipal elections. This figure rises to 34% among left-leaning voters and 46% among supporters of La France Insoumise (LFI).
The implications for the 2026 midterms, and the 2027 presidential election, are significant. As Adélaïde Zulfikarpasic, director of the society division at Ipsos BVA, notes, candidates, media outlets, and policymakers must grapple with the challenges posed by these tools to ensure a transparent and informed democratic process.
A key concern is the lack of transparency surrounding the algorithms that power these AI tools. Questions remain about the sources they utilize, how they select information, and whether inherent biases influence their outputs. The fact that many of the most popular AI tools are foreign-owned – OpenAI’s ChatGPT, Google’s Gemini, Meta AI, and Microsoft Copilot – raises additional concerns about sovereignty and potential external influence.
The rise of AI persuasion also amplifies existing risks to election administration. According to research from Protect Democracy and the AI + Elections Clinic at Arizona State University, AI is “supercharging” long-standing threats such as the malicious spread of false information, intentional cyberattacks on election infrastructure, and inadvertent errors in critical systems. Generative AI provides malicious actors with a powerful new tool to disseminate disinformation at scale, and deepfakes are becoming increasingly difficult to distinguish from authentic content.
The potential for AI to personalize arguments, test their effectiveness, and quietly reshape political views at scale is a particularly worrying development. As one report notes, modern AI doesn’t simply imitate people; it engages in conversations, reads emotions, and tailors its tone to maximize persuasive impact. This capability could enable the creation of coordinated persuasion machines, capable of generating and distributing targeted content without human intervention.
The TIME magazine report suggests that this technological gap is widening, with AI likely to be predominantly used by one political side in the 2026 elections. This asymmetry could further exacerbate existing political divisions and undermine the fairness of the electoral process. The scientificamerican.com article highlights research demonstrating that AI chatbots can measurably sway voters’ choices, even when the information they provide is inaccurate.
As the era of AI persuasion in elections begins, the need for vigilance and proactive measures is paramount. The challenge lies in safeguarding against manipulation, detecting ideological biases in AI algorithms, and ensuring that voters have access to reliable and unbiased information. The upcoming elections will serve as a critical test case for the resilience of democratic institutions in the face of this new technological frontier.
