Research & Papers

New fairness framework tackles allocation of public harms, not goods

Traditional fairness models fail when decisions cause harm. Researchers propose a fix.

Deep Dive

Researchers introduce two new formalizations of the 'core' for fairly allocating public bads—decisions that impose costs on groups. They show that zero-respecting Lindahl equilibria satisfy both core definitions under a structural condition, but without that condition, Lindahl equilibria exhibit undesirable behaviors and sharp impossibility results separate public bads from public goods. A rule using a reduction to public goods recovers one core formalization. The work, accepted to EC'26, lays the groundwork for studying fair allocation of public bads.

Key Points
  • Traditional core fairness fails for public bads; two new formalizations are proposed.
  • Zero-respecting Lindahl equilibria satisfy both fairness criteria under certain structural conditions.
  • Without those conditions, impossibility results arise, but a reduction to public goods recovers one definition.

Why It Matters

Provides a mathematical foundation for fairly distributing societal harms in AI, policy, and resource allocation.

📬 Get the top 10 AI stories daily