This paper explores the trade-offs between fairness and utility in algorithmic decision-making systems that operate over time. Researchers introduced new concepts of group fairness for both immediate and long-term impacts, analyzing how enforcing fairness in the short term can sometimes worsen long-term disparities. The study also demonstrates that simple investment policies can reduce long-term inequalities with a minimal impact on utility, and these findings were validated with both synthetic and real-world data. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Highlights potential long-term fairness issues in sequential decision-making systems, suggesting new policy approaches.
RANK_REASON Academic paper published on arXiv detailing theoretical and empirical analysis of algorithmic fairness. [lever_c_demoted from research: ic=1 ai=1.0]