Researchers have developed and tested privacy-preserving methods for algorithmic shortlisting in participatory budgeting. These algorithms aim to predict which public investment projects are likely to be funded by citizens, using only project features and anonymized historical voting data. A study demonstrated that while a naive approach using a large language model for ranking projects had limitations, a vote-based pipeline allowed state-of-the-art LLMs to achieve performance comparable to classical machine learning methods. The findings suggest that user preferences in participatory budgeting are stable enough for algorithmic shortlisting to effectively approximate an initial project selection. AI
IMPACT Potential to improve efficiency and fairness in citizen-led public investment decisions.
RANK_REASON Academic paper on algorithmic methods for participatory budgeting. [lever_c_demoted from research: ic=1 ai=0.7]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
