Ifop, on behalf of Le Laboratoire de la République, surveyed French citizens about their views on the use of artificial intelligence in the context of the 2027 presidential election. The survey was conducted on a sample of 1,000 people representative of the French population aged 18 and older, from July 16 to 17, 2026.
1 – A Campaign Narrative Dominated by Fear and Distrust
When asked about the images, ideas, and feelings evoked by the use of AI in the 2027 campaign, the vast majority of French people paint a very bleak picture. The primary emotions expressed are fear, rejection, and a loss of trust (35%), followed by misinformation and the blurring of the line between truth and falsehood (18%, of which 14% explicitly cited fake news, lies, and propaganda) and democratic manipulation, interference, and breaches of electoral integrity (15%).
Conversely, positive perceptions—such as usefulness, innovation, modernity, and the view that it is a practical or inevitable tool—account for only 10% of mentions, while the dehumanization and loss of authenticity in political life account for 8%.
Even at the stage of spontaneous associations, electoral AI is thus perceived by the public as a threat far more than as a promise.
2 – More of a Risk Than an Opportunity for Democracy
This sentiment is confirmed when French citizens are asked directly about the nature of AI. Four in ten French citizens (40%) view it primarily as a risk to democracy, while 38% see it as both an opportunity and a risk. Only 5% view it as an opportunity to improve democratic debate.
The perception of risk increases with age, rising from 30% among 18- to 24-year-olds to 47% among those aged 65 and older.
3 – Widespread distrust of AI-generated political news, except among the youngest
The credibility gap is striking: 81% of French people say they do not trust information on political topics generated by artificial intelligence, with 39% saying they “do not trust it at all.” Only 19% say they are confident, with just 2% saying they are “completely” confident.
This distrust, however, masks a significant generational divide: we found a trust rate of 34% among those under 35, compared to just 5% among those aged 65 and older. While older adults are almost unanimously opposed, younger people have a less suspicious attitude toward AI.
4 – Envisioned uses are primarily defensive: seeking information and verifying facts, not delegating
When asked what they would be willing to do with AI in the political sphere, the French describe a use that is primarily informational. The most commonly cited intentions include obtaining explanations of complex policy proposals (44%), understanding the concrete consequences of a reform on their personal situation (43%), detecting misleading or manipulated information (42%), or verifying statements made by political leaders (42%). Comparing candidates’ platforms (41%) and receiving a neutral summary of current events (40%) follow closely behind.
However, none of these uses garners majority support: for each one, a majority of French people answer “no.” And here again, we observe generational differences: among those under 35, these uses are considered by a much higher proportion of respondents (ranging from 50% to 68% depending on the item), compared to only 23% to 30% among those aged 65 and older.
5 – Voting: a red line for a majority of French people
While the French may view AI as an information tool, the vast majority refuse to let it make their choices for them. When presented with an AI system that recommends the candidate whose views most closely align with their own, 68% say they would ignore it entirely, 29% say they would consider it alongside other information, and only 3% say they would likely follow its recommendation.
This reluctance increases with age—82% of those aged 65 and older would simply ignore it—while younger people are more receptive, though they do not rely entirely on AI: 11% of 18- to 24-year-olds would likely follow such a recommendation, and 42% would take it into account. Voting decisions thus remain an area that the French intend to keep under human control.
6 – What the French Will Tolerate from Candidates: AI Is Accepted If It Is Disclosed
When it comes to candidates’ own use of AI, the French rank acceptable practices in order of priority. Producing campaign posters garners the most approval (47%), ahead of analyzing voters’ expectations and concerns (43%), drafting social media posts (38%), and producing campaign videos (38%).
Above all, acceptance remains largely conditional. For the production of posters alone, for example, 7% of the French public accept it without restriction, compared to 40% who accept it on the express condition that its use is disclosed. Transparency thus emerges as the necessary counterpart to tolerance toward AI in campaign contexts.
7 – The Fear of Deepfakes and the Electoral Cost of Openly Using AI
Two findings round out this overview. On the one hand, fear of manipulated content is nearly unanimous: 86% of French people fear that fake videos or recordings created using AI could influence the outcome of the 2027 presidential election, with 42% saying “definitely”—a concern that peaks at 94% among those aged 65 and older.
On the other hand, openly acknowledging the use of AI is seen more as a liability than an asset. If a candidate were to declare that they are using it in their campaign, 45% of French people would be less likely to vote for them and only 6% more likely, with the remaining 49% seeing no impact. This backlash is strongest among younger voters (65% of 18- to 24-year-olds would be put off by it) and on the left (64% among Green Party supporters, 53% among Socialist Party supporters, and 51% among La France Insoumise supporters), while right-wing and National Rally voters remain largely indifferent.
With less than a year to go before the 2027 presidential election, artificial intelligence appears less as a promise of democratic renewal than as a source of mistrust, associated with disinformation, manipulation, and a loss of authenticity in political debate. However, the French do not reject all uses of AI: they may accept it when it helps explain a reform, compare platforms, or verify information. On the other hand, they largely refuse to allow it to influence their voting choices or to replace candidates’ own words, judgment, and accountability. While young people are more willing to use these tools, their attitude remains ambivalent, as they would strongly penalize a candidate who advocates for their use. In this context, transparency, the flagging of user-generated content, and the fight against deepfakes will emerge as essential requirements.
