Melding data, systems, and society
In an era defined by increasingly complex global challenges—from the intricate web of climate change and biodiversity loss to the critical need for regulating advanced artificial intelligence and controlling pandemics—the traditional boundaries of academic disciplines often prove insufficient. While interdisciplinary research is gaining traction, it frequently falls short due to its ad hoc and temporary nature, as highlighted by Munther Dahleh, a distinguished professor of electrical engineering and computer science at MIT.
A decade ago, recognizing this critical gap, Professor Dahleh spearheaded the creation of MIT’s groundbreaking Institute for Data, Systems and Society (IDSS). His vision was to cultivate a more profoundly integrated and enduring collaborative environment, transcending the temporary associations typical of complex problem-solving. This monumental effort and the foundational thinking behind it are now meticulously documented in his new book, “Data, Systems, and Society: Harnessing AI for Societal Good,” published this March by Cambridge University Press.
Professor Dahleh’s book is an attempt to articulate the “driving vision” that led to the institute. It targets various audiences, particularly students aspiring to address pressing societal challenges by leveraging the power of AI and data science. The core message is a call for a more holistic approach to problem-solving, urging future researchers to consider the broader implications of their work.
A central tenet guiding the IDSS’s structure, and a key concept detailed in the book, is what Professor Dahleh refers to as “the triangle.” This framework illustrates the dynamic interplay among three crucial components: physical systems, the people interacting with these systems, and the regulatory and policy frameworks governing them. Data, he explains, forms a vital circle in the middle, connecting all these elements. He emphasizes that when tackling significant societal issues, it is imperative to understand the multifaceted impact of any proposed solution on society, people, and the role of individuals in the system’s success. Neglecting this interaction can lead to solutions that inadvertently marginalize certain groups or fail to achieve their intended societal good.
The COVID-19 pandemic serves as a poignant example of “the triangle” in action. Professor Dahleh points out that the crisis involved the biology of the virus (physical system), the contagion effects driven by social behavior (people), and the governmental decisions on lockdowns and public health measures (regulation/policy). The real-time, complex interaction of these components, often with incomplete data, presented unprecedented challenges. IDSS naturally became a hub for discussing these diverse aspects, underscoring the institute’s foundational premise.
Beyond pandemics, this interconnectedness extends to various modern challenges. Social media and e-commerce platforms, for instance, are prime examples of systems built for human interaction, yet they come with significant regulatory dimensions, especially when addressing issues like misinformation. The book also delves into the critical ethical issues inherent in AI, using self-driving cars as a powerful illustration. While the public might intellectually support programming cars to prioritize innocent lives over the driver’s in dangerous scenarios, this ethical stance often conflicts with practical consumer behavior (people wouldn’t buy such cars), leading to lower adoption rates and, paradoxically, increased casualties overall. This highlights the complex interplay between ethical AI development, human psychology, and policy.
Professor Dahleh clarifies the distinction between “transdisciplinary” and more conventional cross-disciplinary or interdisciplinary research. While all have their merits, he argues that most such efforts are transitory, limiting their long-term societal impact. IDSS was conceived to address this by fostering a lasting structure, offering shared journals, conferences, physical spaces, and a robust sense of community. Its doctoral program, notably, mandates 12 core courses, split equally between statistics, optimization theory, and computation, and the social sciences and humanities. This integrated curriculum is designed to cultivate a culture where students inherently consider all components of “the triangle” simultaneously.
Two years ago, Professor Dahleh stepped down from his leadership role at IDSS to refocus on teaching and research. However, reflecting on the institute’s journey, he realized the absence of a comprehensive document detailing its inception and the underlying thought processes. Unlike his academic research, which is meticulously published, the intellectual journey of creating IDSS remained largely undocumented. This book now fills that void, providing a historical perspective and an accessible guide to how such a pioneering institution came into being. It serves as an invaluable resource for anyone seeking to understand the intricate dance of data, systems, and society, particularly in the age of AI.
