PHIL 4260 Philosophy of AI

Course Overview

The optimistic claims of the early AI researchers ("we'll have an intelligent being in a few decades!") haven't come anywhere close to being true. In fact, many AI researchers have given up on the idea to create a "general-purpose" intelligent being, and instead focus on highly specific techniques to solve highly specific problems. Indeed, any relation between AI and cognitive science can sometimes be hard to find. In this course, however, I want to go back to the original vision of AI: to create an intelligent being, and to inform us about cognition in general. So, in this course, I want to ask the questions: What, if anything, has gone wrong with AI? Do the philosophers of mind have anything to say about this? Can the cognitive psychologists offer some new models that AI could try and implement? And vice versa: what philosophical questions are raised or answered by the successes and failures of AI? And can the cognitive psychologists learn anything from AI? We will talk about general AI techniques such as semantic networks, production systems, neural networks, and natural language processing, and explore what their limitations are in terms of informing us in the study of cognitive phenomena such as reasoning, decision-making, learning, and perception and action.

Connections to Other Courses

There are no official prerequisites for this course, but I do require students to have some general background in philosophy, psychology, computer science, and logic. Excellent background courses would be IHSS-1140 Minds and Machines, COGS-2120 Introduction to Cognitive Science, and of course CSCI-4150 Introduction to AI. Other useful courses are CSCI-1200 Data Structures, CSCI-2300 Introduction to Algorithms, PHIL-2140 Introduction to Logic, PHIL-4140 Intermediate Logic, PHIL-4420 Computability and Logic, and PSYC-4370 Cognitive Psychology.