哈佛教授 David Parkes：全球化生产一支铅笔后，世界的目光看向了 AI | IJCAI 2026

TL;DR - In his IJCAI 2026 John McCarthy Award lecture, Harvard professor David C. Parkes argues that AI’s next frontier is designing rules for societies of specialized agents, rather than relying on a single general-purpose agent. Mechanism design could make decentralized multi-agent systems efficient, fair, incentive-compatible, and robust to strategic behavior.

• Strategy-proof mechanisms encourage agents to report private information truthfully, enabling reliable coordination in auctions, resource allocation, and dynamic environments.
• Computational mechanism design combines economic objectives with practical constraints, including limited preference elicitation and continually changing participants and resources.
• Parkes describes using neural networks and differentiable economics to discover mechanisms while structurally enforcing properties such as strategy-proofness and menu compatibility.
• As AI agents increasingly act on behalf of people or organizations, contract design and moral-hazard controls become important when their actions cannot be directly observed.