Foundations of Opinion Formation in Autonomous Systems

Lead Research Organisation: University of Liverpool
Department Name: Computer Science


The project concerns how groups of partially informed and self-interested agents (e.g., humans, robots), which are faced with a common problem, take a collective decision by exchanging their individual opinions to, possibly, reach a consensus. It aims at understanding how processes of opinion formation in groups behave and how they can be engineered in groups of artificial agents, like robots.

The project capitalizes on techniques developed in the social and economic sciences, applying them to the artificial intelligence setting. It extends the state-of-the-art in the application of voting theory to artificial intelligence, addressing the process of opinion formation, and lays the theoretical groundwork for the development of collective decision-making techniques in autonomous systems.

Planned Impact

The long term vision of the project is to enable advanced collective decision-making capabilities in autonomous systems.

Given the theoretical focus of the project, its short and mid-term impact are likely to be exclusively academic. The project aims at developing a critical mass of theoretical results for a principled understanding of opinion formation in collective decision-making settings involving autonomous systems. The achievement of this aim will be disseminated primarily in AI venues, but venues in the social and economic sciences will also be targeted capitalizing on the support offered by the project partner, the Amsterdam Center for Law and Economics (ACLE).

The mid-term impact of the project will more naturally shift towards applications, even though remaining in an academic context. It will be facilitated by the Center for Autonomous Systems Technology (CAST) of the host institution. The project will benefit from the direct feedback with researchers active in this center, in particular on robotics, and will provide a natural pathway for the project's theoretical findings to benefit research in applications.

Autonomous systems and multi-agent systems are currently a thriving area of research and development. The long-term impact that these systems, in all their incarnations (robots, autonomous vehicles, trading agents, intelligent sensors, and the like) are expected to have on the future of society is vast. The project contributes to their understanding and development, helping to shape their realization.


10 25 50

publication icon
Aucher G (2018) Modal logics of sabotage revisited in Journal of Logic and Computation

publication icon
Christoff Z (2017) Binary Voting with Delegable Proxy: An Analysis of Liquid Democracy in Electronic Proceedings in Theoretical Computer Science

Description We have identified general laws concerning the stabilization of opinion formation on networks, with applications to novel voting systems (e.g., liquid democracy).
Exploitation Route The project carried out foundational research. Our findings have the potential to inform application in autonomous systems technology as well as applications in eDemocracy.
Sectors Digital/Communication/Information Technologies (including Software),Government, Democracy and Justice