I am delighted to participate in today’s forum as a PhD alumnus of the School of Finance. The topic given to me is “Fiscal and Financial Disciplines in the AI Era”. For over two years, global enthusiasm for AI has shown no signs of waning—its momentum remains as strong as ever. Meanwhile, finance is a field we have studied and practiced in for years. So what unfolds when these two intersect? What new realities does this convergence create?
In my view, AI embodies a grand narrative and a sweeping vision of the future. Yet, confronting such a transformative force, humanity can’t help but feel a sense of unease, even insignificance, as if our roles are diminishing. There are already warnings that many people may lose their jobs to AI, calling into question the relevance of much of today’s knowledge. So when this wave of disruption reaches fiscal and financial discipline, how should it redefine its purpose? These questions are sure to come up.
In fact, every industry is constantly being reshaped by technological innovation. Take my current research focus, inclusive finance, as an example. The fintech in this field involves three transformative innovations—AI, internet technology, and smartphones, which have fundamentally reshaped the financial ecosystem. Historically, financial services demanded high-touch, personalized interactions between service providers and clients. Today, such services can be delivered through completely no-touch platforms. This is a clear example of how technological innovation is reshaping the landscape of the financial sector.
But will AI truly render entire disciplines obsolete? As I prepared for this topic, I engaged in deep, though admittedly immature, reflection. I believe we need to properly manage four major relationships.
The first is the relationship between technological advancement and theoretical foundations.
Technology has undeniably taken center stage in today’s discourse. In discussions about AI, buzzwords like machine learning, blockchain, decentralization, big data, algorithmic trading, and quantitative investment dominate the conversation—after all, even DeepSeek emerged from the world of quantitative finance. These technologies appear so formidable, even enigmatic, that they inevitably provoke a question for those of us in the finance sector, and the broader social sciences: in this rapidly evolving landscape, what becomes of our relevance?
Through careful analysis, I’ve come to recognize that technology without theoretical underpinnings lacks the necessary foundation for meaningful application—a potentially dangerous shortcoming. Consider today’s fintech innovations: they all ultimately rest on established financial principles. Without these conceptual foundations, technology becomes like a body with strong limbs but no guiding intelligence. This is precisely why theoretical frameworks remain indispensable. As technological capabilities advance, we must resist the temptation to diminish the enduring importance of financial theory.
The disciplines of finance, investment, and economics continue to demand rigorous theoretical advancement. Much like the nervous system orchestrates bodily functions, theoretical frameworks must guide practical applications. Consider inclusive finance: its foundations rest upon welfare economics, behavioral economics, economics of poverty, and financial sociology—each representing decades of scholarly work. For example, the economics of poverty helps us understand why poverty persists, what defines it at its core, and what the structural mechanisms are that perpetuate it. Without such theoretical depth, no technological solution—however sophisticated its algorithms or processing power—can yield meaningful interventions.
That is why theoretical inquiry remains essential. At the very least, we must strive to strike a balance between technological advancement and theoretical foundations.
The second is the relationship between big data, alternative data, and strong data from field research.
Big data and alternative data have become widely used, and many localities have set up big data bureaus. But does this mean we should stop relying on strong data collected through traditional field research? Can we now rely entirely on machines, with researchers drawing conclusions purely through computer-based analysis, without doing any fieldwork?
Satellite imagery is another example. In the United States, people use satellite images to count the number of cars in shopping mall parking lots as a way to estimate overall sales. In China, MYbank’s Big Dipper system also uses satellite data to measure farmland size and assess loan demand and risk. These cases highlight the growing emphasis on cleaning and analyzing alternative data.
What I wish to emphasize today is that traditional anthropological and sociological methods—particularly in-depth interviews—remain indispensable. Let me illustrate this with an example. Two years ago, CAFI launched the Financial Diaries Project. Our team conducted extensive longitudinal fieldwork, tracking 200 lower-middle-income households across Shanghai, Hunan, and Shaanxi for an entire year. By meticulously documenting their daily income and expenditures over 365 consecutive days, we sought to uncover the root causes behind their financial volatility and, more importantly, to identify the specific types of financial services that would genuinely address their needs.
Our daily tracking revealed a particularly telling case in Hunan. The husband worked as a labor contractor. We observed months where his income suddenly spiked, alternating with periods of unusually high expenditures. While these patterns appeared as mere anomalies in big data visualizations, the true story required human understanding. During my field interview, I asked him what drove these dramatic fluctuations in his finances.
The husband explained that his wife had suffered a fall during Spring Festival, requiring hospitalization which incurred substantial medical expenses. When questioned about the subsequent income rebound in July and August, he revealed this coincided with his high school son’s summer break, during which the teenager contributed to household finances through part-time work.
This case demonstrates the inherent limitations of big data analytics: while exceptionally capable of detecting patterns and anomalies, it cannot interpret the human stories behind each fluctuation. Such understanding still requires field research through anthropological and sociological lenses. Though these methodologies may appear antiquated by today’s technological standards—perhaps even destined to be categorized as “traditional”—their effectiveness remains undeniable. This reality compels us to reconsider any assumptions about phasing out foundational disciplines. Rather than diminishing their relevance, I would argue these fields have never been more critical.
The third is to strike a balance between digital financial services and social financial empowerment.
Digital finance is one of the “Five Priorities” in the financial sector, it is indeed a transformation that has demonstrably enhanced convenience for consumers. Today, virtually all banking services have become accessible through a single smartphone interface.
Yet this digital convenience comes with an ironic trade-off. When contacting bank customer service, we’re frequently greeted by automated systems whose pre-programmed responses often miss the mark of our actual concerns. In these moments, the limitations of technology become painfully clear - what customers truly seek is human connection. This reveals a fundamental truth: financial services are not mere products, but relationships built through genuine interaction. For financial institutions, this serves as a crucial reminder: clients must never be reduced to data points. It’s precisely this understanding that makes social financial empowerment not just valuable, but essential.
Our research on inclusive finance has given us profound insights into this very issue. Take the “last mile” challenge in inclusive financial services—at its core, this requires financial providers to empower clients. It’s not just about delivering financial access, but equipping people with the capability to break free from geographical constraints, whether mountains or deserts. To put it another way: we shouldn’t merely provide blood transfusions (short-term aid); we must help build self-sustaining capacity. This “hematopoietic function” fundamentally means enhancing their long-term financial capabilities.
In practice, financial institutions tend to prioritize product promotion over genuine service and client empowerment. This makes me wonder: are we failing to emphasize this enough in school? I recall lecturing at Renmin University of China—after presenting my inclusive finance theories, students rushed up asking, “Professor Bei, could you tell us more about capital markets instead?” This left me puzzled: why this particular interest?
Students are naturally preoccupied with graduation, employment prospects, and income levels—all legitimate concerns. But this makes me question our educational priorities: are we striking the right balance between teaching digital financial services and fostering social financial empowerment? When we educate future financiers, are we cultivating professionals who instinctively ask, “How can I use my financial skills to empower communities?”
China undoubtedly ranks as a global leader in inclusive finance development. Yet we now face a critical paradox: while achievement metrics excel, financial health indicators are deteriorating alarmingly. Over-indebtedness among vulnerable populations and severe credit impairment have grown so acute that the Report on the Work of the Government called for credit repair. This divergence reveals a fundamental flaw: We’re delivering services but failing at true empowerment. The fundamental issue lies with providers—they’re not transforming financial inclusion beneficiaries into capable, self-sufficient individuals. This represents a critical gap in our current approach.
On this point, I admire the perspective of Professor Robert Shiller at Yale. In his book’ Finance and the Good Society, he contends that structural inequities in society are rooted in flawed financial systems. This struck me as a wake-up call: we’ve always treated finance as merely a tool or instrument, failing to recognize how our daily financial practices inadvertently perpetuate social inequities. So we need to understand that only good finance can help our society achieve equitable prosperity for all.
The fourth is to strike a balance between profit-driven finance and financial ethics.
Our traditional understanding of capital was rather monolithic—we were taught concepts like “if capital can get 100 percent profit, it will ‘trample on all human laws; 300 percent, and there is not a crime at which it will scruple, nor a risk it will not run…” This is what we call “predatory capital”. Does history attest to this? Undoubtedly. But today, we’ve caged such predatory tendencies. More importantly, we now understand that capital comes in many forms. Some of it is greedy, some is purely profit-driven, but some of it has a conscience.
What defines “conscientious capital”? It’s when investments actively weigh social value, environmental protection, and climate impact, demonstrating authentic ethical commitment. Critically, this approach aligns with government advocacy for patient capital, advocating long-termism over short-term gains.
There’s a classic Wall Street saying from Goldman Sachs: “We’re greedy, but long-term greedy”. Note it still uses the word ‘greedy’, but with a far-sighted perspective to avoid short-term pitfalls. To me, patient capital must also be conscientious, not purely profit-driven. It requires both long-term vision and consideration for environmental and social value. Some now advocate for “compassionate capital”, an approach worth noting. Take the Gates Foundation: it is progressively shifting from pure grants to impact investments. Why? To prevent the inefficiency of traditional charity, where funds disappear without a measurable impact. For philanthropic organizations, investing isn’t about high returns—it’s a disciplined mechanism to better achieve their mission.
Policy must impose judicious constraints on predatory and purely profit-driven capital, while vigorously cultivating its conscientious, patient, and compassionate counterparts. Furthermore, our education systems must likewise champion this shift, cultivating financial professionals who instinctively operate through these ethical frameworks.
To sum up, AI will always remain a tool for humans. It can never become Confucius, Aristotle, Keynes, or Marx. Technological revolutions have unfolded for centuries, yet every innovation has ultimately been harnessed by human wisdom. This reality defines our mission in financial education: to cultivate students with deeper theoretical foundations, stronger field research capabilities, broader societal perspectives, and higher ethical literacy. These competencies must be reinforced—not retreated from—in the age of AI.
That concludes the main points. Thank you all.
The End
