In a packed lecture hall at Peking University HSBC Business School, the economist urged a look back through history to understand the intelligence we are building today.

Is artificial intelligence (AI) a 21st-century phenomenon, or are its intellectual origins embedded in the very history of the scientific revolution? This was the provocative premise of Nobel Prize-winning economist Thomas J. Sargent’s public lecture, “AI: Past, Present, and Future,” held at Peking University HSBC Business School (PHBS) on December 29, 2025. To an audience of over 300 faculty, students, and professionals, with thousands more watching online, Professor Sargent masterfully wove threads from cognitive science, the history of science, and economic reasoning to present a deeply historical and human-centric view of AI.

At the event

Professor Sargent began by anchoring the discussion in a clear definition of intelligence itself. He distilled intelligence into three core activities: pattern recognition (compressing streams of sensory data into usable categories), generalization (applying learned patterns to new, unseen situations), and decision-making. AI, in his framing, is simply a machine created by humans to perform one or more of these tasks.

Thomas Sargent gives the lecture

With this lens, Sargent embarked on a historical journey, arguing that the essence of AI—data compression and pattern discovery—is the very engine of the scientific revolution. He portrayed the work of Ptolemy, Copernicus, and Galileo as pioneering acts of “artificial intelligence without computers.” These giants, he explained, systematically “mined” celestial data through painstaking observation, “fitted” geometric and mathematical models to it, and “generalized” their findings into transferable laws.

This historical perspective led to one of the lecture’s most striking contradictions. The core disciplines propelling modern AI—statistics, biology, economics, and physics—are fields where innate human intuition often fails. Yet, it is the convergence of these non-intuitive fields that has catalyzed AI’s greatest leaps. He highlighted AlphaGo as a quintessential example: a system that triumphed by integrating game theory from economics, dynamic programming, and Monte Carlo simulations from statistics and physics.

The climax of the lecture was philosophical. Drawing on physicist Richard Feynman's analogy and philosopher Bertrand Russell's commentary, Professor Sargent posed a fundamental question: Is AI—and indeed science—primarily engaged in “describing” patterns, or in “explaining” the underlying logic and rules governing the world? In a measured tone, he noted: Currently, AI excels at the former, while the latter remains a significant challenge for AI. The central question he leaves us with is not merely technical but profoundly human: As AI grows increasingly adept at describing our world, to what extent can it truly explain it? The future of artificial intelligence may well hinge on bridging this very gap—a challenge rooted in the same spirit of inquiry that has driven scientific discovery for centuries.

Hai Wen delivers remarks

In his remarks, Professor Hai Wen, founding dean of PHBS, vice chairman of Peking University Council, and former vice president of Peking University, summarized the key takeaways from the lecture. He noted that Professor Sargent’s analysis reframed AI not as an entirely new phenomenon, but as the long-term outcome of cross-disciplinary convergence—a process nurtured by the continuous interplay of fields like statistics, biology, physics, and economics. Professor Hai highlighted that while AI demonstrates remarkable strengths in areas like pattern recognition, decision-making, judgment, and information organization, it remains an inadequate tool for proposing entirely new rules or constructing novel cognitive frameworks. Looking forward, he stressed that the path forward must focus on grasping AI's underlying logic and advancing research in core disciplines such as statistics, biology, economics, and physics. By harnessing human creativity and institutional design, he argued, we can continuously propose and shape new rules, thereby steering AI's development more effectively.

Chen Liang in conversation with Thomas Sargent

The lecture sparked immediate and lively engagement from the audience. In a subsequent dialogue moderated by PHBS Assistant Dean and Associate Professor Chen Liang, pressing contemporary questions were addressed, from AI’s impact on labor markets to the essential skills for the AI era. During the Q&A session, Professor Sargent fielded a series of thoughtful questions from the audience. With patience and insight, he addressed how insights from biology and physics could advance AI, what it would take for AI to reason like humans, and how to assess the China-U.S. competition in the field, earning warm applause.

Shi Jiao hosts the event

Associate Professor Shi Jiao, deputy director of the Sargent Institute of Quantitative Economics and Finance at PHBS (SIQEF), served as the host, underscoring the school's focus on global academic engagement. A key driver of this vision since his arrival in 2017, Thomas Sargent helped found SIQEF and has been pivotal in attracting international scholars and enhancing the research ecosystem. He also designed and leads the Economics PhD program, teaching advanced courses including "Advanced Macroeconomics." To date, the institute has admitted seven cohorts of PhD students and organized nine international conferences, creating ongoing exchanges between Chinese researchers and global experts.

Since 2024, PHBS has launched a series of events, for instance, “Dialogue with a Nobel Laureate: PHBS Student Tea Talks." These events regularly invite Professor Sargent to participate in face-to-face lectures, in-depth dialogues, and academic discussions with faculty and students, building an open, forward-looking academic exchange platform with a global perspective.

By Annie Jin and Xing Yaru

Source: PHBS Public Relations and Media Office, SIQFE