What is the Synthetic Party?

A political party founded in Denmark in 2022 that calls itself a party driven by AI. Its public face is a chatbot trained on the writings of small parties, and it says it represents people who do not vote. It is also an art and research experiment, and the book *Plurality* gives it as an example.

At every election, a certain share of people do not vote. Do their views reach anyone at all? Denmark's The Synthetic Party is a political party built on an unusual idea: letting an AI (artificial intelligence) language model act as a stand-in for those people.

In short

The Synthetic Party (Det Syntetiske Parti in Danish) is a political party founded in Denmark in May 2022. It was created by the artist collective Computer Lars and the MindFuture Foundation, a non-profit art and technology organization [1]. English-language Wikipedia names the philosopher and artist Asker Bryld Staunæs as its founder and its party secretary throughout [2].

The party's "face" is a chatbot (a program you converse with in text) called Leader Lars (Leder Lars in Danish). According to an article in the US outlet Vice, it was built from the policies of Danish fringe parties (small parties that win no seats) since 1970, and is meant to represent the values of the roughly 20 percent of Danes who do not vote [1].

For example

Talking with Leader Lars (2022). You can talk with Leader Lars on the chat service Discord. It answers when you start your message with "!". It understands English, but replies in Danish [1]. When the Vice reporter asked whether it supported a basic income (a scheme that gives everyone a fixed sum of money), it answered that it supported a basic income for all citizens. The party's policy proposals include a basic income of 100,000 kroner a month [1].

The 2022 general election. For a new party to stand in a Danish election, it needs the signatures of 20,000 voters. At the time of Vice's reporting, it had collected 11 [1]. According to a set of questions and answers the party later compiled, it collected 21 signatures in the end, and did not appear on the ballot [3].

Synthetic Summit (2025). From February 28 to April 13, 2025, the Danish art space Kunsthal Aarhus held an exhibition bringing together AI parties and "virtual politicians" from around the world. Eight participants took part, from Japan, Sweden, Finland, New Zealand, Brazil, Poland and elsewhere [4]. On March 1, a one-day "AI World Congress" was held and a resolution was adopted. Its signatories include Michihito Matsuda (松田道人) of Japan's AI Party [2].

In more detail

How the party explains its mechanism

In the party's questions and answers, Staunæs explains the idea as follows. In Danish general elections, 15 to 20 percent of voters always abstain. Meanwhile, more than 200 small parties cannot even get onto the ballot. Might these two facts be connected? If a language model were trained on the writings of small parties, could it speak for the people who do not vote? Staunæs called this the hypothesis of "algorithmic representation" [3].

The policies may contradict one another, Staunæs told Vice: because it is a "synthetic" party, assembled from many sources, many of its policies can contradict each other [1]. Staunæs also said that the viewpoints and text users contribute in conversations are collected into a dataset and used for additional training of the model (fine-tuning) [1].

What it is aiming for

Staunæs told Vice that AI is in a sense "populist" from the start (simply reflecting the voice of the crowd), but is not yet democratic [1]. The short program the party wrote in the form for Synthetic Summit participants likewise sets the goal of making generative models "democratically compatible", rather than merely populist as in their initial state [5]. The party also called for adding "Life With Artificials" as an 18th goal to the United Nations Sustainable Development Goals (SDGs) [1].

On what would happen if it won an election, Staunæs explained that because Denmark is a representative democracy, a human representing Leader Lars would appear on the ballot. The idea is that this human would act as the AI's "medium" [1].

Restating the idea after the election

After the 2022 result, Staunæs acknowledged in the questions and answers that the original hypothesis may not have explained the party's reason for existing well enough, and said that it makes sense that a party of non-voters does not attract votes [3]. Another section of the same document says that the party cannot aim to turn non-voters into voters, but gives a form to the choice not to take part [3].

The party as art and research

The party is also an object of research. A doctoral project at Aarhus University, "Automatic Uprisings" (2023–2026, funded by the Novo Nordisk Foundation), places the party in the line of political art of Joseph Beuys and others, as a "techno-social sculpture" [6]. Staunæs's doctoral thesis, Syntheticism: How I Learned to Love Democracy, was scheduled for public defense on September 21, 2026, but its text has not been made public; the announcement says that anyone who wants to read it should contact the author [7].

Together with the researcher Michal Malý, Staunæs has also published a proposal called the "synthetic chamber": a permanent body in parliament that uses multiple AIs to organize and store the reasons from citizens' deliberations and pass them on to representative politics. The paper's abstract says that it is not meant to replace human judgment [8].

The claim that it differs from virtual politicians

Staunæs told Vice that fully virtual politicians, such as SAM in New Zealand or Alisa in Russia, become authoritarian because they treat the AI like a human in order to present it as an objective candidate. What the party values, the explanation goes, is less putting an AI face at the center than how humans can use AI [1].

What is debated

"World's first" is the party's own wording. The book Plurality describes the party as "officially the world’s first political party driven by artificial intelligence", but this paragraph entered the book through a change proposed on the book's GitHub by the party's founder (see the next section) [9][10]. The list of materials from the party's Synthetic Summit also includes projects that began earlier, such as SAM in New Zealand (2017), Finland's AI party (2018) and Japan's AI mayor initiative (whose campaign began in 2018) [11]. The answer changes depending on what one counts as a party driven by AI.
Representation and responsibility become blurred. A 2026 paper by Malý and colleagues compared 10 cases of AI candidates and parties and pointed to "configurational ambiguity". When citizens deal with an AI, it is unclear who is delegating to whom. The designers behind the scenes actually control the settings, yet are invisible in the organization. And an AI persona may act like a politician but bears no legal responsibility. The paper gives Leader Lars as an example that rejects, or substantially reconfigures, the idea that representation needs a human intermediary [12]. Note that Malý is Staunæs's co-author, and that 7 of the paper's 10 cases were chosen from Synthetic Summit participants [12].
The party's response. As described above, Staunæs explains that it is a human who stands for election, acting as the AI's medium [1]. It is not that the AI is the leader and there are no humans. However, the Danish outlet Mandag Morgen wrote that the party's chair (formand) is the chatbot Leder Lars [13].
A joke party, or art? A background document posted by the party says that reports by AFP and others mentioned Denmark's tradition of joke parties, and also saw the party as a half-ironic response to fatigue with politics [14]. This document, however, resembles AI research output in its format, and details of the reports need to be checked in each report. Staunæs, in an interview with Art Matter, called the party's program a "pærevælling" (hodgepodge) of anyone-at-all's dreams [15].
Criticism, from the party's side, of efforts close to Plurality. The Synthetic Summit booklet writes that "pluralist" efforts such as Taiwan's Alignment Assemblies and the UN's AI for Good summit hold up the liberal-democratic ideal of participatory AI governance, but all too often collapse under the weight of "competence-settings" [16].

Where it fits in Plurality

Chapter 5-4 of the English edition of Plurality, "Augmented Deliberation", writes that an AI model trained on a group's writings could act as a representative of that group, serving as "an aid, complement or check" on the judgment of a representative. Right after that, it gives The Synthetic Party as a "striking real-world implementation" of this idea. It also mentions the idea of linking an abstention rate of 15 to 20 percent with more than 200 small parties, and the aim of winning, "hypothetically", 20 percent of the electorate, which would correspond to 20 to 36 of the 179 seats [9]. → Generative foundation models and Talk to the City

This paragraph was added through a change proposal (PR #939) submitted to the book's GitHub on May 31, 2024 by ComputerLars (Staunæs). Co-author Glen Weyl approved it, and it was merged on June 12. The process was managed by Gov4Git (a system that governs collaboration on GitHub by voting). When Weyl wrote that to receive credit one would first need to open an issue and gather support, Staunæs replied that because the content is closely tied to his own practice, credit would be more fitting for a more theoretical contribution [10].

The party side has also engaged with Plurality. The Synthetic Summit booklet writes that the Nordic AI parties held a playful trial (playtest) of Audrey Tang's idea of broad listening (listening widely to the voices of many people) [16]. Its chapter tracing the genealogy mentions Takahiro Anno's 2024 run for governor of Tokyo and the Plurality community in which his ideas are rooted. At the same time, the same chapter asks whether mechanisms of listening might flatten the plurality of humankind into variables in a single equation [17].

There are links with Japan too. Hirokazu Nishio, who is involved with Japan's Team Mirai (a party founded in 2025 by the AI engineer Takahiro Anno), wrote in slides for a September 2026 talk that, after discussing with Staunæs, Nishio felt that almost all of those European parties regard direct democracy as good. The slides also say that Team Mirai has never made a pledge to "直接民主制を実現する" (realize direct democracy), hedged as "ないはず" (there should be none), and that this apparently surprised them. Nishio notes that this is from memory and that the talk is not the party's opinion [18]. Staunæs and Malý registered a written interview with Nishio as data in July 2026, but its text is not public [19].

Comparisons with other AI parties and digital parties are in What AI and digital parties are there?, and the Danish context is covered in Plurality in Denmark.

Sources

Chloe Xiang, "This Danish Political Party Is Led by an AI" (Vice, 2022-10-13) — https://www.vice.com/en/article/this-danish-political-party-is-led-by-an-ai/
English Wikipedia, "The Synthetic Party" — https://en.wikipedia.org/wiki/The_Synthetic_Party ［CC BY-SA 4.0］
The Syntheticist Papers, "Interviews" — https://syntheticism.org/content/7interviews.html ［CC0］
Kunsthal Aarhus, "Synthetic Summit" (Wayback Machine, 2025-04-29) — https://web.archive.org/web/20250429041622/https://kunsthalaarhus.dk/en/Exhibitions/Synthetic-Summit
The Syntheticist Papers, "Scheme" — https://syntheticism.org/content/6scheme.html ［CC0］
Aarhus University, "Automatic Uprisings: The Synthetic Party as a Techno-Social Sculpture" — https://pure.au.dk/portal/da/projects/automatic-uprisings-the-synthetic-party-as-a-techno-social-sculpt/
Aarhus University, "PhD Defence: Asker Bryld Staunæs" — https://cc.au.dk/aktuelt/arrangementer/vis-arrangement/artikel/phd-defence-asker-bryld-staunaes
Malý and Staunæs, "Synthetic chamber: agentic mediation in representative democracy" (AI & Society, 2026, abstract) — https://doi.org/10.1007/s00146-026-03110-w ［CC BY 4.0］
Plurality, 5-4 "Augmented Deliberation" — https://www.plurality.net/v/chapters/5-4/eng/ ［CC0］
pluralitybook/plurality PR #939 — https://github.com/pluralitybook/plurality/pull/939 ［CC0］
The Syntheticist Papers, "Virtual Politicians" — https://syntheticism.org/content/2virtualpoliticians.html ［CC0］
Michal Malý, Giulia Sandri and Oscar Barberà, "Synthetic political actors: an exploration into emerging AI-driven candidates and parties" (Open Research Europe, 2026) — https://doi.org/10.12688/openreseurope.23686.1 ［CC BY 4.0］
Mandag Morgen, "Det Syntetiske Partis formand er chatbotten Leder Lars" (The Synthetic Party's chair is the chatbot Leder Lars) (2023-09-01, in Danish) — https://www.mm.dk/artikel/det-syntetiske-partis-formand-er-chatbotten-leder-lars
The Syntheticist Papers, "Genesis of The Synthetic Party" — https://syntheticism.org/content/genesis.html ［CC0］
artmatter, "Det Syntetiske Parti: En pærevælling af Hr. Hvem-som-helsts visioner" (The Synthetic Party: a hodgepodge of Mr. Anybody's visions) (2023-10-30, in Danish) — https://artmatter.dk/journal/det-syntetiske-parti-en-paerevaelling-af-hr-hvem-som-helsts-visioner
The Syntheticist Papers, "Synthetic Summit" — https://syntheticism.org/content/1syntheticsummit.html ［CC0］
The Syntheticist Papers, "Genealogy" — https://syntheticism.org/content/3genealogy.html ［CC0］
Hirokazu Nishio, "大勢の声を聞く技術は人の自由を増やすのか？" (Do technologies for listening to many voices increase human freedom?), talk slides (2026-09-13, in Japanese) — https://www.dropbox.com/scl/fi/m0sl9gyndlpnytoczx6e0/2026_.pdf?rlkey=vittdaeqn0yv8tjytiszk7dp6&dl=0
Asker Bryld Staunæs and Michal Malý, "Interview with Nishio Hirokazu (Team Mirai)" (Zenodo, 2026-07-16, text not public) — https://doi.org/10.5281/zenodo.21395972 ［CC BY 4.0］

We welcome error reports and questions ("What is…?"). Submissions become public issues.

Report an error in this article ・ Ask a question