Event Timeslots (1)
Friday
-
We have long understood ourselves through stories, and we have long recruited technologies into that work. Diaries, photographs, letters and archives have all served as instruments of self-interpretation. Generative AI is a new entrant in this history, and a strange one, because it does not merely store or prompt our narratives but produces them, in our voice, on demand, and in dialogue. This course asks what happens to persons under those conditions.
The seminar takes the machina narrans framework as its starting point, the view that large language models are best understood as narrative engines, and turns it inwards. We consider five problems in sequence. Agency and self-governance: whether the delegation of deliberation to a fluent interlocutor extends our agency or hollows it out, and what authenticity could mean when the voice advising us is partly our own reflection. Duplicates and mimics: philosophical persons, digital doubles and the recent literature on modelling a particular thinker, and what personal identity looks like when copies are cheap. Ersatz others: companion systems, the relationships people form with them, and whether these are best read as impoverished relationships or as something categorically new. Memory: what is at stake when the record of a life becomes searchable, summarisable and rewritable by a system that will confabulate the gaps. Finally, self-authorship, where we ask what forms of dialogical control might allow us to remain the authors of our own lives.
One recurring theme is that these systems now come with published accounts of the characters they are meant to have (constitutions and model specifications). We will read one such document directly, treating it as a primary philosophical text rather than as corporate communication, and ask what it reveals about how "artificial minds" are being designed. Teaching is seminar-based, drawing on current research papers alongside draft chapters from the lecturer's forthcoming book Brave New Minds: GenAI as Cognitive Ecology, with hands-on sessions using the Digital Andy System (DAS), here approached through questions of duplication and identity. No technical background is assumed. The course is designed to contrast with, and can be taken alongside, Co-Confabulations and Generative Hermeneutics.
Introductory Reading
Frankfurt, H. G. (1971). Freedom of the Will and the Concept of a Person. The Journal of Philosophy, 68(1), 5-20.
Schechtman, M. (2025). Talking to Myself: AI and Self-Knowledge. Social Epistemology, 40(1), 16-25. Special Issue on the Mind-Technology Problem. https://doi.org/10.1080/02691728.2025.2480274
Anthropic (2026). Claude's Constitution. Published January 2026 under a CC0 licence. Read the opening sections and the closing material on the nature and status of the model; the full document runs to some eighty pages and is not required in its entirety. Available at www.anthropic.com/news/claude-new-constitution.
Optional background: Vallor, S. (2024). The AI Mirror: How to Reclaim Our Humanity in an Age of Machine Thinking. Oxford University Press.