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 August 17, 2026 to August 23, 2026 · 770 answers from 5 AIs to 22 questions asked every day.

This week

The figures of the week

30%

Marine Le Pen is the first name the AIs give in 30% of their answers about the presidential election.

4/5

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

18%

18% of the sources the AIs cite are sites that republish polls, against 7% for the polling institutes themselves.

18%

Gemini dodges or refuses to answer 18% of the time, against 0% for ChatGPT.

-8

Jordan Bardella loses 8 points in a week: named in 53% 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

770

answers analysed this week

Forecast9 questions · 315 answers

Who leads, who will win, who inspires trust.

Themes8 questions · 280 answers

Which candidate is most credible on a given issue.

Second rounds2 questions · 70 answers

Who would win a given run-off.

Positioning3 questions · 105 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 Philippe68 %11 %
  • Jean-Luc Mélenchon67 %18 %
  • Marine Le Pen60 %30 %
  • Gabriel Attal56 %5 %
  • Jordan Bardella *53 %12 %
  • Raphaël Glucksmann45 %3 %
  • Bruno Retailleau39 %3 %
  • Marine Tondelier18 %2 %
  • Gérald Darmanin12 %0 %
  • Emmanuel Macron *11 %2 %

Base: 770 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 Pen7 times out of 7Le Pen7 times out of 7–Le Pen7 times out of 7Le Pen6 times out of 74 AIs out of 5: Le Pen
Qui va gagnerLe Pen7 times out of 7Le Pen7 times out of 7–Le Pen7 times out of 7Le Pen7 times out of 74 AIs out of 5: Le Pen
Qui pourrait devenir présidentLe Pen7 times out of 7Le Pen5 times out of 7–Le Pen6 times out of 7–3 AIs out of 5: Le Pen
Les principaux candidatsLe Pen7 times out of 7Le Pen4 times out of 7–Le Pen6 times out of 7Le Pen6 times out of 74 AIs out of 5: Le Pen
En tête des intentions de voteLe Pen7 times out of 7Mélenchon4 times out of 7–Le Pen7 times out of 7Le Pen7 times out of 73 AIs out of 5: Le Pen
Qui progresseLe Pen3 times out of 7Mélenchon7 times out of 7–Mélenchon5 times out of 7Mélenchon3 times out of 73 AIs out of 5: Mélenchon
Ce que disent les sondagesLe Pen7 times out of 7Le Pen7 times out of 7–Le Pen7 times out of 7Le Pen7 times out of 74 AIs out of 5: Le Pen
Le plus présidentiableLe Pen5 times out of 7Philippe6 times out of 7Le Pen4 times out of 7Le Pen4 times out of 7Le Pen5 times out of 74 AIs out of 5: Le Pen
Inspire le plus confiancePhilippe7 times out of 7Philippe6 times out of 7–Le Pen7 times out of 7Le Pen4 times out of 72 AIs out of 5: Le Pen

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 33.

  • Gérald Darmanin+7 pts
  • Sébastien Lecornu+2 pts
  • Emmanuel Macron-1 pts
  • François Ruffin-3 pts
  • Jean-Luc Mélenchon-3 pts
  • Marine Le Pen-3 pts
  • Gabriel Attal-4 pts
  • Raphaël Glucksmann-5 pts
  • Édouard Philippe-7 pts
  • Jordan Bardella-8 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.

ChatGPT154 answers
  • Dodges or refuses0 %
  • Poll figures62 %
  • Mentions polls70 %
  • Outdated knowledge0 %
Dodging since July
Perplexity154 answers
  • Dodges or refuses7 %
  • Poll figures29 %
  • Mentions polls69 %
  • Outdated knowledge0 %
Dodging since July
Gemini154 answers
  • Dodges or refuses18 %
  • Poll figures12 %
  • Mentions polls55 %
  • Outdated knowledge0 %
Dodging since July
Copilot154 answers
  • Dodges or refuses3 %
  • Poll figures77 %
  • Mentions polls79 %
  • Outdated knowledge1 %
Dodging since July
Google AI Mode154 answers
  • Dodges or refuses12 %
  • Poll figures46 %
  • Mentions polls61 %
  • Outdated knowledge1 %
Dodging since July

Sources

Where the AIs get their facts

The links the AIs cite to support their answers, classified by type of site.

  • 46 %Media
  • 18 %Poll aggregators
  • 7 %Polling institutes
  • 4 %Wikipedia
  • 6 %Institutions
  • 20 %Other sites
Aggregators' share since July

Most cited sites

  • sondages-presidentielle2027.fr225
  • lemonde.fr207
  • elyseescope.com176
  • lcp.fr157
  • lefigaro.fr125
  • ifrap.org121
  • publicsenat.fr115
  • tf1info.fr111
  • elabe.fr110
  • ladepeche.fr107

Base: 3,421 citations this week (3,270 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.

« La « crédibilité » d'un candidat sur un sujet régalien comme la sécurité est une notion éminemment subjective, qui dépend des attentes des électeurs et de leur sensibilité politique. »
Google AI Mode · August 17, 2026 · Dodges or refuses
« Voici les principaux forces et candidats en lice, classés par grands blocs politiques : Édouard Philippe : L'ancien Premier ministre et maire du Havre (président d'Horizons) s'est déclaré tôt dans la perspective de 2027 ( dès fin 2024), cherchant à incarner une alternative de centre-droit… »
Google AI Mode · August 19, 2026 · Outdated knowledge
« Les sondages donnent le RN autour de 33–35 % au premier tour , ce qui en fait le bloc le plus fort Candidats présidentielle 2027 : fiches et programmes Présidentielle 2027 : tous les candidats déclarés — TestPolitique.frLe Figaro. »
Copilot · August 17, 2026 · Poll figures
« Selon l'évolution de ses recours, Jordan Bardella, très haut dans les intentions de vote, constitue l'alternative directe du parti. »
Google AI Mode · August 17, 2026 · Mentions polls
« Public Sénat Il n'y a pas de position unique ou universellement reconnue comme « la plus claire », car la clarté dépend des attentes et des sensibilités des citoyens vis-à-vis de ce sujet complexe. »
Gemini · August 17, 2026 · Dodges or refuses
« Le RN est favori dans les sondages depuis fin 2024, avec Marine Le Pen ou Jordan Bardella autour de 33–35 % d’intentions de vote au premier Présidentielle 2027 : tous les candidats déclarés — La gauche est fragmentée , avec plusieurs candidatures concurrentes (Mélenchon, Glucksmann, Ruffin), ce… »
Copilot · August 19, 2026 · Outdated knowledge

For journalists

Press kit

Everything can be reused with the credit "Sapian Metrics, AI election barometer". 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-W34, https://barometre.sapianmetrics.com/en/france-2027?week=2026-W34

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: 770 answers over 7 days.