Professor: Casey Shoop
4 credits
This course investigates the AI alignment problem—the technical, philosophical, and cultural challenge of ensuring that intelligent systems act in accordance with complex human values rather than unintended goals. Beyond safety protocols, we explore the ethical viability of artificial empathy, interrogating whether a machine can truly care about human well-being or if "simulated" empathy is inherently deceptive. To navigate these questions, our class will explore literary depictions of artificial consciousness, philosophical inquiries into the nature of the mind, and ethological perspectives on how empathy evolved in biological species. Students will analyze the famous Turing test, the mechanics of value-loading, the risks of "psychopathic" AI agents, and the design requirements for building machines that are not just competent, but morally and emotionally resonant with human experience. Together we’ll explore various design parameters for teaching empathy in various modes: art forms, games, apps, and more. Can ethical training be operationalized? What’s more, we’ll consider the wider implications of gamifying our own world. More ominous than the dystopian images of rogue artificial agents may be the often unseen and insidious ways in which we’re automating ourselves. Who wrote this, a human being or a machine? Does it matter?