To what extent can an artificial system be rational?
New MIT course, 6.S044/24.S00 (AI and Rationality) does not attempt to answer this question. Instead, it challenges students to explore this and other philosophical issues through the lens of artificial intelligence research. For the next generation of scientists, concepts of rationality and agency may prove to be integral to AI decision-making, especially if it influences how humans understand their own cognitive limitations and limited, subjective views of what is and is not rational.
This study is rooted in the deep connections between computer science and philosophy, which have long collaborated in formalizing what it means to form rational beliefs, learn from experiences, and make rational decisions in pursuit of one's goals.
“You might imagine that computer science and philosophy are quite far apart, but they have always intersected. The technical parts of philosophy really overlap with artificial intelligence, especially early artificial intelligence,” says course instructor Leslie Kaelbling, a Panasonic professor of computer science and engineering at MIT, recalling Alan Turing, who was both a computer scientist and a philosopher. Kaelbling herself holds a bachelor's degree in philosophy from Stanford University and notes that computer science was not available as a field of study at the time.
Brian Hedden, a professor in the Department of Linguistics and Philosophy who holds a joint appointment in the MIT Schwarzman College of Computing with the Department of Electrical Engineering and Computer Science (EECS), who teaches Kaelbling, notes that the two disciplines are more related than you might think, adding that “the differences are a matter of emphasis and perspective.”
Tools for further theoretical thinkingG
First presented in fall 2025, Kaelbling and Hedden created artificial intelligence and rationality as part of the project Common ground for IT education, a cross-cutting initiative of the MIT Schwarzman College of Computing that brings together multiple departments to develop and teach new courses and launch new programs that combine computer science with other disciplines.
With over twenty students registered, AI and Rationality is one of two Common Ground classes with a philosophical foundation, the other being 6.C40/24.C40 (Computer Science Ethics).
While Computer Ethics examines concerns about the social impacts of rapidly evolving technology, Artificial Intelligence and Rationality examines the contested definition of rationality by considering several elements: the nature of rational agency, the concept of a fully autonomous and intelligent agent, and the attribution of beliefs and desires to these systems.
Because artificial intelligence has incredibly broad applications and each use case presents different problems, Kaelbling and Hedden brainstormed topics that could provide fruitful discussion and engagement between the two perspectives of computer science and philosophy.
“When I work with students studying machine learning or robotics, it's important for them to step back a little and check their assumptions,” Kaelbling says. “Thinking about things from a philosophical perspective helps people better understand how to place their work in a real-world context.”
Both instructors emphasize that this is not a course that provides concrete answers to questions about what it means to design a rational agent.
Hedden says, “I see this course as building a foundation. We're not giving them a set of doctrines to learn, memorize, and then apply. We're equipping them with the tools to think critically about things as they enter their chosen career, whether they work in research, industry, or government.”
The rapid advancement of artificial intelligence is also creating new challenges in academia. Predicting what students will need in five years is an impossible task for Kaelbling. “We need to give them higher-level tools — habits and ways of thinking — that will help them approach things that we really can't predict right now,” he says.
Connecting disciplines and challenging assumptions
So far, the class has attracted students from a wide range of disciplines, from those firmly rooted in computer science to others interested in exploring how artificial intelligence connects to their own fields of study.
Throughout the semester of reading and discussion, students grappled with different definitions of rationality and the ways in which they contradicted the assumptions of their fields.
Amanda Paredes Rioboo, a senior at EECS, talks about what surprised her about the course: “We were kind of taught that math and logic are the gold standard or truth. In this class, they showed us many examples of people behaving inconsistently with this mathematical and logical framework. We opened up a whole can of worms to see if it's people who are irrational? Is it the machine learning systems we designed that are irrational? Is it the math?” and logic itself?”
Junior Okoroafor, a doctoral student in the Department of Brain and Cognitive Sciences, appreciated the challenges facing the class and the possibilities of changing the definition of a rational entity depending on the discipline. “Presenting what each field means to rationality in a formal framework clarifies precisely which assumptions are intended to be shared and which differ across fields.”
The co-teaching and collaborative course structure, as with all Common Ground endeavors, gave students and instructors the opportunity to hear different perspectives in real time.
For Paredes Rioboo, this is her third Common Ground course. He says, “I really like the interdisciplinary aspect. I always thought it was a nice mix of theory and application because they have to cross fields.”
According to Okoroafor, Kaelbling and Hedden showed obvious synergy between the disciplines, saying it seemed like they were engaging and learning with the class. The way computer science and philosophy can be used to inform each other allowed him to understand their similarities and invaluable perspectives on intersecting issues.
He adds: “philosophy can also surprise.”
