The Understanding of Machines - A Language-Focused Introduction to the Philosophy of AI

Course Overview

In this course, we will explore both the subjectivus and the objectivus reading of »the understanding of machines«, addressing two central questions: (1) How do we determine whether a machine truly understands a language, and (2) what does it mean to understand machines of the kind used and created in contemporary AI applications. Both questions are predominantly (but not exclusively) conceptual questions, centering around the meaning of certain key concepts. The course is introductory in nature: It is intended both for undergraduate philosophy students as well as for students from other disciplines with an affinity for the topic.

The course has a clear focus on language on two levels. On the technical level, this implies a focus on AI systems that process natural language, such as large language models (LLMs). On the conceptual level, this implies a focus on the concept of understanding, as it is applicable in understanding claims, concepts, etc. that are phrased in natural languages.

We start out by exploring what it means to understand language, building on the basic pragmatic insight that language is essentially a social practice as well as on the specific contributions of Robert Brandom’s inferential pragmatism. Next, we familiarize ourselves with important insights from the philosophy of AI regarding the understanding that we can develop of LLMs. Here, the emphasis lies on the classics in the field, such as the introduction by Russell and Norvig and John Searle’s Chinese Room Thought Experiment. We then proceed to a theoretical and practical introduction to LLMs.

Building on these three parts, we can then address the two guiding questions of the seminar in its fourth and final part. It is dedicated to cutting-edge discussions that (1) attempt to chart the linguistic understanding of LLMs as well as (2) research in the philosophy of explainable AI that attempts to help us understand these machines.

University Course Directory: Link