arXiv:2508.06577v3 Announce Type: replace-cross Abstract: Participatory budgeting is a democratic innovation that allows citizens to propose and vote on public investment projects. To help organizers manage large volumes of submissions, we design and test privacy-preserving methods for algorithmic shortlisting. These algorithms predict which projects are likely to be funded using only project features and anonymous historical voting data. We demonstrate the limitations of a naive approach that uses a large language model to rank projects based on past success and propose a vote-based pipeline
The increasing volume of citizen submissions in participatory budgeting initiatives necessitates automated solutions, while advancements in AI, specifically large language models, offer new tools for managing and analyzing this data.
This development indicates a growing application of AI in direct democratic processes, potentially streamlining governance and enhancing public engagement while raising questions about algorithmic bias and fairness in decision-making.
Algorithms are now being designed to influence democratic processes by pre-selecting public investment projects, shifting how citizen input is processed and prioritized.
- · Government organizers of participatory budgeting
- · AI algorithm developers
- · Citizens whose projects are prioritized efficiently
- · Citizens whose projects are unfairly deprioritized
- · Traditional manual project review processes
- · Organizations focused on purely human-driven governance
More efficient and scalable participatory budgeting processes are implemented in cities and municipalities globally.
Public trust and engagement in democratic processes may increase due to perceived efficiency, or decline due to concerns about algorithmic transparency and fairness.
The success of these algorithms could lead to their broader adoption in other forms of public policy and decision-making, further integrating AI into governance structures.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI