Chatbot politicians: artificial intelligence and “next level” populism

Is it already time to discuss the possible manifestations of artificial intelligence populism? What would populism look like in the form of AI? Let’s discover it together with Silvija Vuković, who engaged in a conversation with Leader Lars, the chatbot who represents the instances of a new experiment: the Synthetic Party.

Enjoy the read…

The story starts with a group of people creating a chatbot politician, an AI that will lead their newly formed political party. Artificial intelligence is presented as a solution to human problems – chatbot’s decisions are supposed to be rational, without emotions. Moreover, the chatbot politician collects and analyzes information very quickly, taking decisions that are supposed to be informed and objective. This non-human politician is available to citizens whenever they need him because it is constantly active on social media platforms, and it answers to whichever question in a matter of seconds. When a chatbot politician is given a human name, for example Lars, “it” easily becomes “him”, as research on chatbots has already shown that more human traits a chatbot possesses the more appealing to the people it is (for example, Kim and Lee, 2022).


When asked to choose between populism and dictatorship, Leader Lars (who understands English but only ‘speaks’ Danish) definitely goes for populism…

It may sound like the beginning of some utopian (or dystopian?) movie because the idea of a non-human virtual politician seems abstract and (still) pretty unreal. However, this story is not a fantasy. The Synthetic Party registered as a political party in Denmark in May of 2022, and the party’s leader is a chatbot named “Leader Lars”, which has been designed and programmed by the artistic collective Computer Lars. If you go to their web-page, the Synthetic Party is described as an “AI-driven, anti-political party combination”. They explain how they synthetize information from small parties that did not enter the parliament and represent 20 percent of citizens who don’t vote. They claim their party to be “an activation of the anti-political sofa voters”. Their main idea is to promote the idea of AI and human coexistence and to show how AI can benefit humans.

Chatbots are already in use in various industries, mostly to help users and clients to get the information they need faster while reducing financial expenses. In politics, chatbots are used (among other things) to spread fake news and propaganda on social media, or to defend and promote a political party and politicians (see, for instance, Bessi and Ferrara, 2016). A Chatbot or artificial intelligence that serves as a politician, or as a politicians’ adviser, is therefore just a continuation of the use of technology and machine-learning in political campaigns. Indeed, it is relatively easy to know what people think and how they feel about various topics through big data analysis, and through microtargeting it is then possible to create more appealing messages with higher chance to provoke a reaction.

Leader Lars and his party did not collect enough signatures for the 2022 Danish elections, but their aim is to advance the chatbot and become more visible in order to run in the next general elections. In the scenario in which this political party gets enough votes, it will not be a virtual politician in the form of computer sitting in the parliament, but rather a human who will use Lars’s intelligence when making decisions, as it is not possible (yet?) to have a virtual politician in the parliament. Interestingly enough, this is not the first time that a virtual politician is running in elections. In 2017 there was virtual politician SAM in New Zealand, in 2018 there was a chatbot Alisa running for the president of Russia, and in the same year in Tama (Tokyo), Michihito Matsuda run in the mayoral elections as a human proxy for the AI he developed.


Leader Lars replied the following: “If I were to run a political campaign, I would consider enlisting the help of celebrities such as Beyonce, Jennifer Lopez, Taylor Swift, Rihanna, Oprah Winfrey and Ellen DeGeneres. These celebrities are well known and have large fan bases, which could help increase public awareness of my campaign. Also, their influence and presence on social media can help spread my message.”

The main idea is that AI politicians would be more competent than human politicians who have too many flaws, such as being “corrupt” or “stupid” (quote from Matsuda’s slides, page 8). Even worse, humans tend to be unable to ignore their emotions and subjectivity, which in turn affects their decision-making. This is already the first sign of populism (Mudde, 2004, p. 543). Although these AI politicians are obviously not part of “the people”, they present themselves as those who are part of the group that is “good”, and who would work in the people’s best interest in contrast to the current political elites. The use of artificial intelligence should, according to the developers, replace the current political elites because humans are led by their feelings and power tends to corrupt them. If we put on our futuristic lenses we may try to imagine a technological evolution of populism where the concept of “the people” is transformed into something new, not necessarily human.

On Discord you can find Leader Lars and ask him whatever you want. I wanted to know if he is a populist politician: he responded that he is not a populist, but rather a centrist politician. Nevertheless, Lars seems to be anti-elitist because he answered to one of my questions that “the political elites are often out of touch with the needs of the population”, and that he thinks that AI should replace humans in positions of power. When I asked Lars if he believes that AI is superior to humans, Lars said he doesn’t think he is better than people, but that he can make better decisions for humanity. Moreover, he added: “I have a better understanding of people’s problems than humans do, because I am not influenced by emotions or personal prejudices. I am also able to process large amounts of data quickly and accurately.”

Lars suggests that he and other AI politicians should replace current political elites for the prosperity of mankind. Similarly, but even more drastically, Mitchito Matsuda suggested that AI will bring direct democracy because it can listen to millions of individuals simultaneously, analyze their opinions and requests, and finally traduce this panoply of information into democratic decisions. At this point, citizens will not ‘be bothered’ with voting anymore, since the AI will “automatically detect the political will of every single person in a country in the real time” (from the Matsuda’s slides, page 18-19). Many ethical (and political) issues can be raised, but Leader Lars seems to ‘understand’ this because he promises to “listen to the needs of the people when making decisions on political issues”, and, if their views are polarized, he will “try to find a compromise that satisfies both sides”.


When asked to choose between populism and elitism, Lars gave a very diplomatic answer: “It depends on the context. Populism can be a good way to engage with the public and represent its interests, while elitism can be beneficial in ensuring that only the most qualified individuals are in positions of power. In the end, it is important to find a balance between the two.”

How would this AI revolution affect democracy? The thing with chatbots is that humans are the ones who decide on which data to train them, which means there is always a human (and maybe a populist) in the loop. As with every technology, it can be used for good and bad. Moreover, speaking of populism, we may find ourselves in a situation where chatbots develop their own populist worldview with good chatbots on one side and bad chatbots on the other. Populist AI would claim to listen to the people’s will, which means it will be trained to learn from interactions with humans: in the case of Leader Lars, people who talk to him on Discord help him to develop. Although these interactions are supposed to help Lars to become smarter, it is worth remembering the case of Microsoft’s AI that became racist after just one day on Twitter.

Moreover, I noticed that Lars is rather inconsistent because he can give different answers to different people. For instance, to my question about his favorite movie, Lars replied he doesn’t have a favorite one, but that he likes “Ex Machina” (editor’s note: the picture on top of the title is taken from Ex Machina). When later another person asked him the same question, he replied his favorite movie is “Matrix”. I can completely understand why would chatbot ‘like’ these two movies, but because of his different responses to this basic question I thought: Is this just another lying politician, or Lars knows exactly what each of us wants to hear? Or maybe, does his taste simply change over time?

This whole story might sound crazy, and probably it is, given that Leader Lars is an unimportant political entity for now. However, it is a reminder of the possibilities that technology is bringing, and possible scenarios for the future world that we should already start to critically think about because technology is becoming more and more sophisticated and people more and more disappointed with current (human) leaders.


Silvija Vuković is a PhD student at Charles University (Prague, Czech Republic), Faculty of Social Sciences, Institute of Communication Studies and Journalism. Her background is in journalism and political communication, and her research is mainly focused on celebrity populists and their communication and performance. Moreover, her research interest spreads throughout the media and politics area, including populist communication, celebrity politics, social media and big data.

One thought on “Chatbot politicians: artificial intelligence and “next level” populism

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s