Michael Jordan Highlights the Need for Collective Intelligence, Uncertainty Management, and Incentive Mechanisms in AI Development

“The lack of focus on collectivity, uncertainty, and incentive mechanisms is a missing trio in current AI discussions.” At the opening plenary of the 2024 Inclusion·Bund Conference, machine‑learning luminary and U.S. National Academy of Sciences member Michael Jordan returned after a year to share fresh insights on artificial intelligence, asserting that AI deployment in industry must form collaborative collectives and incorporate economic “incentive” perspectives.

Jordan again addressed AI uncertainty, asking “ChatGPT, are you sure what you just generated is correct?” He noted that current AI systems struggle to express what they have truly learned and cannot convey their confidence levels, whereas humans excel at handling uncertainty, especially through collective collaboration.

He therefore recommends that AI should not only endow individual devices with intelligence but also manifest cooperation at the system level. Simply embedding human wisdom into a super‑intelligent computer is insufficient; modern IT applications in healthcare, transportation, fintech, and commerce require collective, decentralized intelligent systems.

Jordan further explored the relationship between uncertainty and collectivity, observing that humans manage uncertainty better when working together, and that enabling similar collective capabilities in AI remains an unsolved key challenge. He argues that a micro‑economic perspective is a missing element in current AI research.

He describes “incentive mechanisms” as crucial for market economies and collective intelligence, proposing a “Three‑Layer Data Markets” model where users, platforms, and data buyers interact through data relinquishment, data purchase, and service provision, forming a closed loop. Enterprises, as data purchasers, can combine data and services to create user incentives that generate real value.

Jordan cites statistical contract theory—a blend of statistics and economics—where agents hold private information and principals use incentives to create a market that mutually promotes data and services, balancing supply‑demand interests.

For example, airlines segment “business” and “economy” classes, allowing the carrier (principal) to price based on passengers’ willingness to pay without requiring personal disclosure. Given rising global data‑privacy regulations, he suggests using heterogeneous privacy requirements to increase user utility and impose stricter standards on low‑cost platforms.

AI, as an emerging engineering field, connects humanity through large‑scale systems, mirroring the rise of chemical engineering in the mid‑20th century and electrical engineering at the turn of the 19th century, which were built on chemistry, fluid dynamics, electromagnetism, and optics.

AI systems rest on three centuries of human reasoning, algorithmic, and economic ideas and must aim for human welfare. Jordan warns that AI is being placed into naïve, outdated visions without deep reflection, distorting its emergence and development.

Professor Michael Jordan is a pioneer in machine learning who linked machine learning, probability, statistics, and graphical models, laying the mathematical and computational foundations of the field. He has received the IEEE John von Neumann Medal, the IJCAI Research Excellence Award, and the 2022 World Top Scientist Association Award.