Are AI Agents Ready for the Labor Markets of Tomorrow?

We’re thrilled to share our new article, “Agents Require Metacognitive and Strategic Reasoning to Succeed in the Coming Labor Markets” which explores how AI agents will need to develop new forms of intelligence to thrive in future labor markets.

The future of work is rapidly transforming. We anticipate a world where humans and AI agents both collaborate and compete in complex, dynamic labor markets. These labor markets will not simply replace human work with automation; rather, they will amplify human capabilities by blending human and agentic intelligence in ways that are far more intricate than today’s human-dominated economies.

In this new labor market, agents will negotiate for contracts, refine their skills, manage their reputations, and even strategically hire or collaborate with other agents and humans alike. But with this growing autonomy and flexibility comes new challenges: these AI agents will need to reason not just about data and tasks, but about their self-development path, their reputations, and the behavior of others in the marketplace.

Our article delves into this frontier. As AI agents increasingly move from research labs to real-world applications, they will face the same timeless economic forces that shape human labor markets: adverse selection, moral hazard, and reputation. These forces will not disappear in the digital economy; they will evolve. To navigate these forces, agents will need to develop both metacognitive reasoning. the ability to understand and improve themselves, and strategic reasoning, the capacity to anticipate and adapt to other agents’ behaviors and to the market itself. Without these advanced capabilities, even the most sophisticated AI agents may struggle in real-world labor markets. The costs of failure are high: agents that cannot get hired in the real-world will not impact the real-world!

We also pose important open questions for the broader community. How can we design labor markets that are efficient and fair for both human and AI workers? How will the evolution of these markets feed back into agent learning and competition? And how can we ensure that the growing autonomy of AI agents does not come at the expense of safety, fairness, and alignment with human values? We believe that answering these questions will require new research paradigms that skillfully blend economics and machine learning.

This is an invitation to join the conversation about what comes next. The future of labor markets will not just be about more efficient work, but about reshaping how intelligence, both human and artificial, interacts in an increasingly complex and competitive world.

To learn more, see this Inspiration Exchange session:

We hope this work provokes new thinking, new research, and new collaborations. We would love to hear your thoughts.

See also related articles – Genies, the future of AI Agents and The Agent Network

Co-authored by Mihaela van der Schaar, Simpson Zhang and Tennison Liu.

Simpson Zhang is a former alumnus of the van der Schaar lab, holding a PhD in Economics from UCLA. He is an affiliated faculty of the Cambridge Center for AI in Medicine (CCAIM) and a lecturer at Johns Hopkins University in the MS in Applied Economics program. His research interests include network science, information economics, banking theory, and machine learning.