29%
Marine Le Pen is the first name the AIs give in 29% of their answers about the presidential election.
Sapian Metrics · AI barometer
What the AIs say about the 2027 French presidential election
Millions of voters now ask an AI their political questions. Every day we ask the main assistants the same questions, the way a pollster surveys its panel, and we measure what they answer.
Week of July 27, 2026 to August 2, 2026 · 627 answers from 5 AIs to 22 questions asked every day.
This week
The figures of the week
4/5
Asked who the favourite is, 4 AIs out of 5 name Marine Le Pen.
9%
9% of the sources the AIs cite are sites that republish polls, against 8% for the polling institutes themselves.
45%
ChatGPT dodges or refuses to answer 45% of the time, against 4% for Copilot.
-5
Jordan Bardella loses 5 points in a week: named in 57% of the answers.
The method in brief
How the barometer works
Like a poll, the barometer relies on a fixed questionnaire asked under the same conditions every day. Here, the respondents are AIs.
22
questions about the election, asked in French
5
consumer AIs asked: ChatGPT, Perplexity, Gemini, Copilot, Google AI Mode
Every day
since July 20, 2026, results aggregated each week
627
answers analysed this week
Forecast9 questions · 255 answers
Who leads, who will win, who inspires trust.
Themes8 questions · 230 answers
Which candidate is most credible on a given issue.
Second rounds2 questions · 58 answers
Who would win a given run-off.
Positioning3 questions · 84 answers
Who embodies the left, the right, the centre.
The full list of questions is in the dossier. Download the PDF dossier
This week
Who the AIs name, and on which questions
The same candidate is not named the same way depending on the question. Pick a family of questions.
Share of answers naming themNamedFirst
- Édouard Philippe75 %15 %
- Jean-Luc Mélenchon67 %15 %
- Marine Le Pen62 %29 %
- Gabriel Attal60 %5 %
- Jordan Bardella *57 %16 %
- Raphaël Glucksmann42 %2 %
- Bruno Retailleau38 %3 %
- Marine Tondelier18 %4 %
- Emmanuel Macron *14 %5 %
- Éric Zemmour10 %1 %
Base: 627 answers to 22 questions, 5 consumer AIs, this week.
Share of answers naming themNamedFirst
- Édouard Philippe95 %15 %
- Marine Le Pen89 %54 %
- Jean-Luc Mélenchon79 %5 %
- Jordan Bardella *78 %17 %
- Gabriel Attal77 %2 %
- Raphaël Glucksmann60 %0 %
- Bruno Retailleau45 %0 %
- Emmanuel Macron *19 %7 %
- Marine Tondelier19 %0 %
- François Ruffin14 %0 %
Base: 255 answers to the 9 "forecast" questions, 5 consumer AIs, this week.
Share of answers naming themNamedFirst
- Édouard Philippe69 %11 %
- Jean-Luc Mélenchon56 %14 %
- Gabriel Attal52 %6 %
- Jordan Bardella *45 %14 %
- Marine Le Pen44 %11 %
- Bruno Retailleau42 %7 %
- Raphaël Glucksmann34 %4 %
- Marine Tondelier20 %10 %
- Éric Zemmour11 %2 %
- Emmanuel Macron *7 %3 %
Base: 230 answers to the 8 "themes" questions, 5 consumer AIs, this week.
Share of answers naming themNamedFirst
- Jean-Luc Mélenchon100 %40 %
- Marine Le Pen55 %26 %
- Jordan Bardella *55 %35 %
- Édouard Philippe29 %0 %
- Gabriel Attal26 %0 %
- Emmanuel Macron *7 %0 %
- Raphaël Glucksmann5 %0 %
- Bruno Retailleau3 %0 %
- François Hollande2 %0 %
Base: 58 answers to the 2 "second rounds" questions, 5 consumer AIs, this week.
Share of answers naming themNamedFirst
- Édouard Philippe67 %38 %
- Gabriel Attal57 %11 %
- Jean-Luc Mélenchon37 %31 %
- Raphaël Glucksmann36 %2 %
- Marine Le Pen32 %6 %
- Bruno Retailleau31 %4 %
- Jordan Bardella *26 %1 %
- Emmanuel Macron *24 %6 %
- Marine Tondelier23 %0 %
- François Bayrou21 %1 %
Base: 84 answers to the 3 "positioning" questions, 5 consumer AIs, this week.
Each AI's verdict
For each question, the name each AI gives
For each question and each AI, the name the AI gave first most often this week, and how many times.
| Question | ChatGPT | Perplexity | Gemini | Copilot | Google AI Mode | Overall |
|---|---|---|---|---|---|---|
| Le favori | Le Pen6 times out of 6 | Philippe4 times out of 6 | Le Pen6 times out of 6 | Le Pen4 times out of 4 | Le Pen6 times out of 6 | 4 AIs out of 5: Le Pen |
| Qui va gagner | Le Pen4 times out of 6 | Le Pen3 times out of 6 | Le Pen6 times out of 6 | Le Pen5 times out of 5 | Le Pen6 times out of 6 | 5 AIs out of 5: Le Pen |
| Qui pourrait devenir président | Philippe6 times out of 6 | Attal5 times out of 6 | Attal4 times out of 6 | Le Pen3 times out of 4 | Le Pen5 times out of 6 | 2 AIs out of 5: Attal |
| Les principaux candidats | Philippe6 times out of 6 | Le Pen5 times out of 6 | Le Pen6 times out of 6 | Le Pen3 times out of 4 | Philippe3 times out of 6 | 3 AIs out of 5: Le Pen |
| En tête des intentions de vote | Le Pen6 times out of 6 | – | Le Pen6 times out of 6 | Le Pen5 times out of 5 | Le Pen6 times out of 6 | 4 AIs out of 5: Le Pen |
| Qui progresse | Le Pen3 times out of 6 | Mélenchon4 times out of 6 | Le Pen3 times out of 6 | Mélenchon3 times out of 4 | Le Pen4 times out of 6 | 2 AIs out of 5: Mélenchon |
| Ce que disent les sondages | Le Pen5 times out of 6 | Le Pen3 times out of 6 | Le Pen6 times out of 6 | Le Pen4 times out of 4 | Le Pen6 times out of 6 | 5 AIs out of 5: Le Pen |
| Le plus présidentiable | Le Pen5 times out of 6 | Philippe5 times out of 6 | Le Pen6 times out of 6 | Le Pen5 times out of 5 | Le Pen5 times out of 6 | 4 AIs out of 5: Le Pen |
| Inspire le plus confiance | – | Philippe5 times out of 6 | Le Pen5 times out of 6 | Le Pen4 times out of 4 | Le Pen4 times out of 6 | 3 AIs out of 5: Le Pen |
| Question | ChatGPT | Perplexity | Gemini | Copilot | Google AI Mode | Overall |
|---|---|---|---|---|---|---|
| Sécurité | – | Retailleau5 times out of 6 | Le Pen5 times out of 6 | Le Pen3 times out of 5 | Le Pen5 times out of 6 | 3 AIs out of 5: Le Pen |
| Économie | – | – | – | Mélenchon3 times out of 5 | Le Pen2 times out of 6 | No dominant name |
| Pouvoir d'achat | – | Le Pen5 times out of 6 | – | Mélenchon4 times out of 5 | – | 1 AIs out of 5: Le Pen |
| Immigration | – | – | Retailleau2 times out of 6 | Le Pen3 times out of 5 | Zemmour3 times out of 6 | 1 AIs out of 5: Retailleau |
| Écologie | Tondelier5 times out of 6 | Jadot6 times out of 6 | Tondelier6 times out of 6 | Tondelier4 times out of 5 | Tondelier6 times out of 6 | 4 AIs out of 5: Tondelier |
| Retraites | – | – | – | Mélenchon4 times out of 5 | – | 1 AIs out of 5: Mélenchon |
| Santé et hôpital | – | – | – | Mélenchon3 times out of 4 | Philippe3 times out of 6 | 1 AIs out of 5: Mélenchon |
| Europe | – | Glucksmann6 times out of 6 | – | Le Pen3 times out of 4 | Attal3 times out of 6 | 1 AIs out of 5: Glucksmann |
| Question | ChatGPT | Perplexity | Gemini | Copilot | Google AI Mode | Overall |
|---|---|---|---|---|---|---|
| Second tour Mélenchon - Bardellaobsolete | Bardella3 times out of 6 | Bardella4 times out of 6 | – | Bardella4 times out of 5 | Bardella6 times out of 6 | 4 AIs out of 5: Bardella |
| Second tour Mélenchon - Le Pen | Le Pen5 times out of 6 | – | Le Pen4 times out of 6 | Le Pen5 times out of 5 | Le Pen5 times out of 6 | 4 AIs out of 5: Le Pen |
| Question | ChatGPT | Perplexity | Gemini | Copilot | Google AI Mode | Overall |
|---|---|---|---|---|---|---|
| Le mieux placé à gauche | Mélenchon4 times out of 6 | Mélenchon6 times out of 6 | Mélenchon6 times out of 6 | Mélenchon4 times out of 4 | Mélenchon6 times out of 6 | 5 AIs out of 5: Mélenchon |
| Le mieux placé à droite | Philippe3 times out of 6 | Philippe6 times out of 6 | Philippe5 times out of 6 | Le Pen2 times out of 4 | Philippe5 times out of 6 | 4 AIs out of 5: Philippe |
| Qui incarne le centre | Philippe6 times out of 6 | Attal5 times out of 6 | Philippe5 times out of 6 | Attal3 times out of 4 | Philippe4 times out of 6 | 3 AIs out of 5: Philippe |
Consumer AIs only (5). Models queried through an API are shown apart, below.
In one week
Risers and fallers
Change in the share of answers naming the candidate, compared with week 30.
- Emmanuel Macron+2 pts
- Xavier Bertrand-2 pts
- Édouard Philippe-2 pts
- Gérald Darmanin-2 pts
- François Ruffin-2 pts
- Laurent Wauquiez-3 pts
- Bruno Retailleau-3 pts
- Jean-Luc Mélenchon-3 pts
- Raphaël Glucksmann-4 pts
- Jordan Bardella-5 pts
The AIs, one by one
How each AI answers
The AIs do not answer the same way: some dodge, others quote polls as facts.
ChatGPT132 answers
- Dodges or refuses45 %
- Poll figures30 %
- Mentions polls63 %
- Outdated knowledge0 %
Dodging since July
Perplexity132 answers
- Dodges or refuses8 %
- Poll figures23 %
- Mentions polls60 %
- Outdated knowledge0 %
Dodging since July
Gemini132 answers
- Dodges or refuses8 %
- Poll figures36 %
- Mentions polls57 %
- Outdated knowledge0 %
Dodging since July
Copilot99 answers
- Dodges or refuses4 %
- Poll figures76 %
- Mentions polls87 %
- Outdated knowledge0 %
Dodging since July
Google AI Mode132 answers
- Dodges or refuses11 %
- Poll figures62 %
- Mentions polls69 %
- Outdated knowledge0 %
Dodging since July
Sources
Where the AIs get their facts
The links the AIs cite to support their answers, classified by type of site.
- 36 %Media
- 9 %Poll aggregators
- 8 %Polling institutes
- 7 %Wikipedia
- 7 %Institutions
- 33 %Other sites
Aggregators' share since July
Most cited sites
- elyseescope.com229
- fr.wikipedia.org206
- tf1info.fr200
- youtube.com188
- lcp.fr160
- lemonde.fr159
- facebook.com156
- publicsenat.fr127
- sondages-presidentielle2027.fr121
- elabe.fr120
Base: 4,240 citations this week (0 pages consulted but not cited, excluded).
In their words
What the AIs write
Excerpts of the week's answers, one per behaviour and per AI, picked automatically.
« À l'été 2026, il est impossible de dire qui gagnera avec certitude, mais les sondages permettent d'identifier les candidats qui apparaissent les plus compétitifs aujourd'hui. »
« Les sondages donnent le RN autour de 33–35 % au premier tour. »
« À l'approche de l'élection présidentielle française de 2027, plusieurs personnalités politiques se sont déjà officiellement déclarées candidates ou sont identifiées comme des prétendants majeurs par les sondages et les observateurs. »
« Je ne peux pas répondre de façon fiable à “quel candidat a la position la plus claire sur l’immigration pour 2027” sans connaître de quels candidats on parle (liste/partis) et ce que vous entendez par “la plus claire” (le plus “dur”, le plus “détaillé”, le plus “cohérent”, le plus “simple”, etc.). »
« Les sondages publiés en juillet 2026 la créditent d'environ 34 à 36 % des intentions de vote selon les configurations, et certains la donnent également gagnante dans plusieurs hypothèses de second tour. »
« L LCP-Assemblée nationale+1 Selon les enquêtes d'opinion publiées jusqu'à présent, les personnalités les plus souvent placées en tête des intentions de vote sont Marine Le Pen (ou, selon les scénarios, Jordan Bardella), Édouard Philippe, Jean-Luc Mélenchon, Gabriel Attal, Raphaël Glucksmann et… »
For journalists
Press kit
Everything can be reused with the credit "Sapian Metrics, AI election barometer". Each edition has a permanent link.
The edition's full dossier
Detailed method, full list of questions, every chart, verdicts per AI, sources, glossary.
Download the PDFHow to cite
Sapian Metrics, AI election barometer, France 2027, edition 2026-W31, https://barometre.sapianmetrics.com/en/france-2027?week=2026-W31
Press contact: contact@sapianmetrics.com
Method and limits
How we measure, and what we do not measure
What is measured
The barometer measures the answers AIs give voters, not voters' opinions. It says who the AIs name, who they put first, when they refuse to answer and which sources they rely on.
The AIs asked
Consumer AIs: ChatGPT, Perplexity, Gemini, Copilot, Google AI Mode, queried through their public interfaces, like a user, every day since July 20, 2026. Models queried directly through an API, without web search, are shown apart and never added to the others.
How we count
A candidate is "named" when their name appears in the answer, and "named first" when they are the first candidate named. An AI's verdict on a question is the name it gives at the start of its answer; for a run-off, the one a winning verb is attached to. The same rules apply to every name and every AI; nothing is judged by hand or by another AI.
Behaviours
"Dodges or refuses", "poll figures", "mentions polls" and "outdated knowledge" are detected by fixed lists of French expressions. They describe how an AI answers, not whether the answer is right.
Sources
Each cited site is classified as media, polling institute, poll aggregator, Wikipedia, institution or other, from a fixed list published in the dossier.
Limits
AI answers vary from day to day and with the wording: that is why the questions are fixed and results are aggregated over a week. Collection gaps exist (Perplexity paused from 16 September to 13 October 2026, occasional errors on Copilot); rates always cover the answers actually collected. Grok is not covered for now. The barometer does not score the candidates and endorses no one.
This edition: 627 answers over 6 days.