In Memoriam: Brian Cantwell Smith (1950 - 2025)
Former dean wove philosophy and computation into a lifelong inquiry into the human dimensions of intelligence, judgment, and meaning
Born into a prominent Canadian family known for both its intellectual achievements and record of public service, UC faculty member Brian Cantwell Smith found his calling in the philosophical foundations of computing and artificial intelligence, emerging fields where he could shape the future while staying true to the family traditions of scholarship and commitment to the public interest.
Smith held the University of Toronto’s Reid Hoffman Chair in Artificial Intelligence and the Human from 2018 until his retirement in August 2025. A former dean of the Faculty of Information, he had appointments in Information, Philosophy, Computer Science, Cognitive Science, and the History and Philosophy of Science and Technology. In his role as Dean, Smith drew on philosophy, computer science, and psychology to redefine what an Information School could be. His vision helped transform a small, library-focused Faculty into a dynamic, interdisciplinary centre. He also brought the Faculty of Information into iSchools, a thriving international organization of information schools of which it is still an active member.
Smith’s course, “Minds and Machines,” inspired many undergraduates who went on to make their mark in philosophy, tech, and science studies. Among the students he mentored was a young Reid Hoffman, who helped build PayPal and co-founded LinkedIn. Decades later, when Hoffman made an early investment in OpenAI, which runs ChatGPT, he got back in touch with Smith.
“I started thinking about the philosophy again, not just the practicalities of building businesses and products and services and investment and all, but what does AI mean for humanity,” Hoffman said.
In their conversations, Hoffman remembers telling Smith, “The entire world needs to grapple more deeply with these things you’ve been thinking about your entire life.” Those discussions led to the creation of the Reid Hoffman Chair in Artificial Intelligence and the Human at U of T. Prior to coming to U of T, Smith was Professor of Cognitive Science, Computer Science, Philosophy, and Informatics at Indiana University, and, after that, Kimberly J. Jenkins University Distinguished Professor of Philosophy and New Technologies at Duke University. His research focused on the philosophical foundations of computing, artificial intelligence, and mind, as well as on fundamental issues in metaphysics and epistemology.
Born in Montreal in 1950, Smith grew up in Canada, India and the U.S. surrounded by family members who were renowned in their fields. His father Wilfred, his uncle Arnold Cantwell Smith, and his brother Julian Smith were all recipients of the Order of Canada, the country’s highest honour. Wilfred, a renowned scholar of Islam and comparative religions who developed and led centres for world religions at McGill and Harvard Universities, was recognized for his influence on religious studies and enhancing the relationships between religious communities. Arnold, a diplomat, was recognized for his service as ambassador to Egypt and the Soviet Union and as the first Commonwealth Secretary-General. Julian, a restoration architect, was recognized for his time as chief restoration architect at Parks Canada, where he was instrumental in establishing a program to protect federally owned heritage buildings.
On his maternal side, Smith’s grandfather, Gordon Struthers, was a medical missionary in China, who, upon returning home, joined the Ontario Department of Health where he established the first Health Unit Services. Smith’s mother Muriel became a physician in Lahore when it was still part of India.
Through his philosophical work on the ontology and epistemology of AI systems, Smith continued his family’s contributions to society. His 2019 book, The Promise of Artificial Intelligence: Reckoning and Judgment (MIT Press) was reviewed in both The Times Literary Supplement and The London Review of Books, as well as countless scholarly journals. It garnered public attention because it foresaw many of today’s ongoing debates about the capabilities and lack thereof of Artificial General Intelligence. Smith argued that AI would be unlikely to become capable of a particularly human quality involved in intelligence, which he called “judgment” and defined as a form of “deliberative thought, grounded in ethical commitment and responsible action, appropriate to the situation in which it is deployed.”
Murray Shanahan, Professor of Cognitive Robotics at Imperial College, London, said of the book, “Brian Cantwell Smith’s timely book presents a philosophically nuanced account of the profound conceptual obstacles that still need to be overcome if we are ever to endow machines with human-level general intelligence.”
In his book, Smith contrasted this ability to judge with reckoning, which, in the philosophical sense of the word, means calculating. While he acknowledged the startling, calculation-based results that can be achieved with current machine learning and large language models, he maintained that judgment is not simply a way of processing the world, but that it emerges from a particular relationship to the world that machines do not have. A system with judgment must, Smith insisted, not simply be able to process but also to “care about what it is processing.” To echo Smith’s close friend and colleague, the late philosopher John Haugeland, it must “give a damn.”
On this, Smith parted company from another good friend, the Nobel laureate and Professor Emeritus of Computer Science at U of T, Geoffrey Hinton. “We completely disagreed on the nature of intelligence. And we had a lot of fun arguing about it,” said Hinton. “I think part of the disagreement was whether the current chatbots really understand what they’re saying. I think they really do understand what they’re saying, and they understand it in the same way as we understand it. We had very friendly arguments because Brian was incredibly intellectually generous.”
Smith first began thinking seriously about AI back in the seventies (now branded as Good Old-Fashioned Artificial Intelligence or GOFAI), when it was still science fiction to most. After studying briefly at Oberlin College and working as a computer programmer at the National Research Council in Ottawa, Canada, Smith found his way to MIT. His talent was quickly recognized by the newly appointed head of MIT’s Artificial Intelligence Laboratory, Patrick Winston. Keen to admit Smith to the graduate program, Winston gave him an informal oral exam covering topics from the undergraduate curriculum. Based on his performance, Smith was awarded an MIT Bachelor’s degree allowing him to matriculate at MIT for his Master’s and PhD.
Smith’s dissertation, Procedural Reflection in Programming Languages, developed the concept of a reflective programming language in what he called 3-LISP, a dialect of LISP (List Processing), which was, at the time, the primary programming language for AI research. Through the development of 3-LISP, Smith worked out how a computer could reflect about itself, its own operations, and the world in which it was engaged. At heart, it was a study of the philosophical foundations of meaning and intelligence, grounded in the technical structure of the 3-LISP programming language.
While completing his doctoral dissertation, Smith was invited to work at the then, groundbreaking multidisciplinary, XEROX Palo Alto Research Center (PARC) where he would ultimately be hired as Principal Scientist. His good friend and colleague from PARC, Lancaster University Professor Emeritus Lucy Suchman, said of his dissertation, “It was quite technical and very well regarded, but once he finished his PhD, I think he became progressively more engaged with working in cognitive science and philosophy of mind. Brian was always a philosopher at heart.”
Thanks to PARC’s close ties to Stanford University, Smith found fertile ground for his philosophical interests. There, he deepened his unique perspective on the epistemology and ontology of computing systems through his friendships and philosophical conversations on computing with John Barwise, mathematician and logician (1942-2000); Geoffrey Nunberg, linguist (1945-2020); and John Haugeland, philosopher of mind and Heidegger scholar (1945-2010) – all of whom became life-long friends and intellectual companions.
Of Smith’s first book, On the Origin of Objects, Haugeland wrote, “This is an essay in fundamental metaphysics, but not like any we’ve ever seen before. Bringing to ontology the training of a computer scientist, and the sensibilities of an artist-engineer, Smith recreates our understanding of objects essentially from scratch – and changes, I think, everything.”
Through all his work – scholarship, teaching, program development, and administration – Smith remained true to his family’s commitment both to spirituality and moral grounding in the physical world. This distinct perspective was also reflected in his love of nature and fondness for machines, mechanisms, and making things.
On his island, Amanda, in Georgian Bay, Ontario, Smith introduced Hinton, a frequent visitor, to power pullers. “We once demolished a shed using a power puller,” said Hinton. “You can pull things with a force of two tons. We both loved weird mechanical devices.”
To get to the island and back, Smith was inspired to buy a trawler which he named Simpliciter – a philosophical term meaning, simply or without qualification – and then a Ford 350 pick-up to haul Simpliciter to water. Even during emergency transport to the hospital, Smith admired the complex machinery of ambulances. His wife Gillian Einstein remarked at his funeral that, had he not been sick, Smith would have loved to have taken apart the hospital machines that were keeping him alive.
Smith had a sense of fun that infused his scholarship, teaching, and administrative work. His many published papers and talks often had double entendres for titles. For example, he called a paper posing the question of whether computers are truly digital, “Indiscrete Affairs”, and another, exploring the relationship between science and God (dedicated to his father), “God, Approximately”. He named the strategic plan in which he laid out the future of the Faculty of Information, “Stepping Up”, and rather than referring to it as a white paper, he called it “Chartreuse”.
Smith’s final book project, Computational Reflections, now in production at MIT Press, continues his lifelong conversation on the philosophical foundations of thought, intelligence, and humanity, especially in computation where the human stakes have become so high.
Brian Cantwell Smith is survived by his wife, Gillian Einstein, his stepson, Alexander Gopen, daughter-in-law, Briana Cowie, his siblings Arnold, Julian, Heather, and Rosemary as well as many beloved nieces and nephews.
A Festschrift celebration of Brian Cantwell Smith will be held on February 9, 2026 at the University of Toronto. For more information, contact communications.ischool@utoronto.ca.