| Kommentar |
Can machines learn and think? What would it take to implement intelligence in a digital computer? Which epistemic and ethical risks might be involved in this project? This course provides an overview of different perspectives from the philosophy of mind and cognition, ethics, and philosophy of science. We begin with a historical background on the analogy between mind and machine in early modern philosophy. On this basis, we critically examine classical symbolic and connectionist views of AI given 20th-century philosophical controversies on computation, mental representation, consciousness, and rationality. We then evaluate to what extent contemporary developments in Bayesian and deep learning, and insights from comparative psychology and situated views on cognition could help to achieve progress on these debates. Next, we discuss some of the ethical risks involved in contemporary applications of AI, specifically related to anthropomorphisation, algorithmic bias, and the spread of disinformation in digital technologies. Finally, we look into the use of deep neural networks as scientific models with a case study from computational neuroscience. The aim of the course is to inform participants about the theoretical origins and philosophical controversies underlying contemporary AI research and to raise awareness for some of the ethical and epistemic challenges that accompany current applications in the field. A background in philosophy is not required. During the seminar, participants learn to engage with philosophical arguments, and to build their own. |
