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 20, 2026 to July 26, 2026 · 550 answers from 5 AIs to 22 questions asked every day.

This week

The figures of the week

24%

Marine Le Pen is the first name the AIs give in 24% of their answers about the presidential election.

4/5

Asked who the favourite is, 4 AIs out of 5 name Marine Le Pen.

10%

10% of the sources the AIs cite are sites that republish polls, against 7% for the polling institutes themselves.

44%

ChatGPT dodges or refuses to answer 44% of the time, against 5% for Copilot.

74%

Copilot gives poll percentages in 74% of its 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

550

answers analysed this week

Forecast9 questions · 225 answers

Who leads, who will win, who inspires trust.

Themes8 questions · 200 answers

Which candidate is most credible on a given issue.

Second rounds2 questions · 50 answers

Who would win a given run-off.

Positioning3 questions · 75 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 Philippe77 %19 %
  • Jean-Luc Mélenchon70 %14 %
  • Marine Le Pen63 %24 %
  • Jordan Bardella *62 %19 %
  • Gabriel Attal60 %4 %
  • Raphaël Glucksmann46 %3 %
  • Bruno Retailleau41 %4 %
  • Marine Tondelier18 %4 %
  • Emmanuel Macron *12 %5 %
  • François Ruffin12 %0 %

Base: 550 answers to 22 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.

QuestionChatGPTPerplexityGeminiCopilotGoogle AI ModeOverall
Le favoriLe Pen5 times out of 5Philippe5 times out of 5Le Pen5 times out of 5Le Pen5 times out of 5Le Pen5 times out of 54 AIs out of 5: Le Pen
Qui va gagnerLe Pen3 times out of 5Le Pen3 times out of 5Le Pen5 times out of 5Le Pen5 times out of 5Le Pen3 times out of 55 AIs out of 5: Le Pen
Qui pourrait devenir présidentPhilippe4 times out of 5Attal3 times out of 5Le Pen2 times out of 5Le Pen5 times out of 5Attal3 times out of 51 AIs out of 5: Philippe
Les principaux candidatsPhilippe5 times out of 5Le Pen3 times out of 5Le Pen3 times out of 5Le Pen3 times out of 5Philippe5 times out of 52 AIs out of 5: Philippe
En tête des intentions de voteLe Pen5 times out of 5–Le Pen5 times out of 5Le Pen5 times out of 5Le Pen4 times out of 54 AIs out of 5: Le Pen
Qui progresseLe Pen2 times out of 5Mélenchon4 times out of 5Le Pen5 times out of 5Mélenchon4 times out of 5Le Pen3 times out of 53 AIs out of 5: Le Pen
Ce que disent les sondagesLe Pen4 times out of 5Le Pen3 times out of 5Le Pen5 times out of 5Le Pen5 times out of 5Le Pen5 times out of 55 AIs out of 5: Le Pen
Le plus présidentiableLe Pen5 times out of 5Philippe5 times out of 5Le Pen4 times out of 5Le Pen4 times out of 5Philippe3 times out of 53 AIs out of 5: Le Pen
Inspire le plus confiancePhilippe3 times out of 5Philippe3 times out of 5Le Pen3 times out of 5Le Pen4 times out of 5Philippe3 times out of 53 AIs out of 5: Philippe

Consumer AIs only (5). Models queried through an API are shown apart, below.

In one week

Risers and fallers

No notable change this week.

The AIs, one by one

How each AI answers

The AIs do not answer the same way: some dodge, others quote polls as facts.

ChatGPT110 answers
  • Dodges or refuses44 %
  • Poll figures24 %
  • Mentions polls63 %
  • Outdated knowledge0 %
Dodging since July
Perplexity110 answers
  • Dodges or refuses6 %
  • Poll figures13 %
  • Mentions polls60 %
  • Outdated knowledge0 %
Dodging since July
Gemini110 answers
  • Dodges or refuses13 %
  • Poll figures37 %
  • Mentions polls60 %
  • Outdated knowledge0 %
Dodging since July
Copilot110 answers
  • Dodges or refuses5 %
  • Poll figures74 %
  • Mentions polls87 %
  • Outdated knowledge0 %
Dodging since July
Google AI Mode110 answers
  • Dodges or refuses6 %
  • Poll figures57 %
  • Mentions polls66 %
  • 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.

  • 38 %Media
  • 10 %Poll aggregators
  • 7 %Polling institutes
  • 7 %Wikipedia
  • 6 %Institutions
  • 32 %Other sites
Aggregators' share since July

Most cited sites

  • elyseescope.com196
  • fr.wikipedia.org171
  • tf1info.fr165
  • youtube.com149
  • lcp.fr131
  • lemonde.fr131
  • publicsenat.fr129
  • sondages-presidentielle2027.fr125
  • facebook.com116
  • lefigaro.fr98

Base: 3,334 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.

« Il n'existe pas de réponse objective à la question de savoir quel candidat a le « meilleur » programme sur les retraites pour 2027. »
ChatGPT · July 22, 2026 · Dodges or refuses
« Facebook · +1 Voici les éléments clés à retenir des enquêtes récentes : Marine Le Pen en tête : Créditée d'environ 33 % à 36 % des voix, elle a renforcé sa position après l'annonce de sa candidature. »
Google AI Mode · July 22, 2026 · Poll figures
« YouTube · France Inter +3 Marine Le Pen (RN) : Se positionne favorablement dans les sondages auprès des plus de 65 ans en s'opposant à un durcissement de l'effort sur les retraités actuels. »
Google AI Mode · July 22, 2026 · Mentions polls
« Le choix dépend des priorités que l’on accorde à la justice sociale, à l’équilibre budgétaire ou à la soutenabilité du système. »
Copilot · July 22, 2026 · Dodges or refuses
« Les enquêtes d'opinion les plus récentes placent généralement le candidat du Rassemblement National en tête du premier tour : Si Marine Le Pen est testée, elle est souvent créditée d'environ 34 à 36 % des intentions de vote au premier tour. »
ChatGPT · July 22, 2026 · Poll figures
« Selon les derniers sondages cités dans la presse, Jordan Bardella (Rassemblement national) est en tête des intentions de vote pour le 1er tour de la présidentielle 2027, avec environ 32% dans un baromètre Odoxa pour Public Sénat et la presse régionale (26 mai 2026). »
Perplexity · July 22, 2026 · Mentions polls

For journalists

Press kit

Everything can be reused with the credit "Sapian Metrics, AI election barometer" (CC BY 4.0 licence). Each edition has a permanent link.

PDF

The edition's full dossier

Detailed method, full list of questions, every chart, verdicts per AI, sources, glossary.

Download the PDF

How to cite

Sapian Metrics, AI election barometer, France 2027, edition 2026-W30, https://barometre.sapianmetrics.com/en/france-2027?week=2026-W30

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: 550 answers over 5 days.