Presidential 2022: developed in Toulouse, a platform allows citizens to build their program

With less than a fortnight before the first round of the 2022 presidential election, a team of French, American and Chilean scientists has imagined and developed a digital participation platform that allows users to create their own program.

The tool was designed and developed by a team from the Center for Collective Learning, an interdisciplinary research laboratory of the 3IA Institute of the University of Toulouse (ANITI), in collaboration with lawyers, computer scientists and mathematicians, including experts in social choice theory from the Computer Science Research Institute of Toulouse (IRIT), the CNRS, the Paris 3IA PRAIRIE institute and the Paris-Dauphine University. The team was led by Cesar Hidalgo, an American-Chilean researcher specializing in artificial intelligence, former director of the prestigious MIT in the United States who came to settle in France in 2020, in particular to work within the institute of artificial intelligence of Toulouse.

Toulouse attracts an American star of artificial intelligence

Myprogram2022 is a collaborative exercise that allows users to build their own political platform by ranking more than 100 proposals extracted from the programs of the main candidates for the 2022 presidential election. The designers of the platform have included the proposals of the 12 candidates, but mainly from those who obtain more than 2% in polls (Macron, Le Pen, Mélenchon, Zemmour, Pecresse, Jadot, Roussel, Lasalle and Hidalgo).

Identify areas of agreement/disagreement with other participants

How does it work ? The online tool asks citizens to select the proposals they agree with and then rank them in order of preference. By selecting and ranking proposals, users help create both individual and collaborative programs.

Users can choose proposals according to the themes: Employment & Economy, Environment and Energy, Public Finances & Debts, Taxation, Civil Service, Governance & Republic, Interior Defense & Justice, Public Policies Social & Solidarity, Territories & Communities. Among the latter, it is possible to find, for example: the increase in the minimum wage to 1,400 euros net per month, the end of the 35-hour week, the privatization of audiovisual public service, the establishment of full autonomy for Corsica or even the lowering of the age of criminal majority from 18 to 16 years.

Monprogramme2022 also allows them to identify points of agreement and disagreement with the other participants.


“Steering the debate towards the most controversial issues”

Accessible to the public since March 29, 10 a.m. on, the platform is not intended to influence the presidential election. Through it, the researchers want to “learn more about how to design scalable and engaging digital engagement platforms” and better understand the empirical limits of social choice theory.

My2022program presents different modes of participation that scientists hope will allow them to develop better tools and learn the mathematical structure of individual and collective preferences. The research group still hopes that their tool “may influence the conversation surrounding“the presidential election without affecting it.

“If our platform is widely adopted, it can steer the debate towards the most controversial issues instead of consensus ones (where no debate is needed). But the platform does not seek to influence elections, but to instantly measure the preferences of the population on 120 subjects a few days before the poll”, indicate the researchers behind the project.

“There is a growing consensus that democracy has serious problems, but there is disagreement on how to fix them. Fortunately, experiences like MyProgram help teach us how to do it,” says Umberto Grandi , senior researcher on the project and computer science lecturer at Toulouse Capitole University and IRIT.

“There is a moral imperative to see if we can use technology to modernize our institutions. We have learned in previous research that these platforms can be effective in identifying divisive issues,” adds César Hidalgo, director of the Center for Collective Learning. , whose role was to lead the project’s UI design, vision and strategy.

Use the data for future projects and research

At ANITI, the question of algorithmic influences is important. This is why one of the goals of the Monprogramme2022 platform is to help generate data that can help researchers think about how to build better recommender systems. Thus, after the presidential election, the data collected on the platform will be used to test “empirically certain ideas of social choice theory” and to learn how to design clear and engaging participation systems.

AI: Vitesco and the Aniti Institute join forces for a more advanced industry 4.0

“We will use the data to understand how to better design digital engagement platforms and to test ideas from social choice theory. Regarding the first objective, we are interested in user behavior. How long do they stay on the site? Do they come back? How many preferences do they cater to? Where are they confused? Designing a digital participation system that is “perfect” in theory, but that people don’t understand or use, is a failure in practice With respect to social choice theory, we will use the collected data to understand how to estimate agreement and disagreement, and to explore matrix formulations of classical social choice theory concepts,” the group of researchers explain.