Inaugural lectures mark not merely an appointment, but a commitment — a scholar's public declaration of the questions that animate their work and the intellectual terrain they intend to open. It is with great pleasure that the Faculty of Arts and Humanities at King's College London invites you to the inaugural lecture of Professor Mercedes Bunz, Professor in Digital Cultures and Societies, Department of Digital Humanities.

Professor Bunz's scholarship occupies that rare and demanding position at the intersection of critical theory, media studies, and the technical realities of contemporary AI systems — bringing to that encounter the conceptual precision of the humanities, her deep curiosity for the technical, and a refusal of easy answers.

In her lecture ‘Culture as Calculation’, Professor Bunz takes up one of the most consequential intellectual challenges of our moment. Drawing on contemporary AI safety research from leading AI labs, she argues that the language produced by large language models is not simply human language reproduced by machines. It is something altogether more interesting. If the humanities in the twentieth century analysed symbolic culture, she contends, their task in the twenty-first century is to understand vector culture: to understand how meaning emerges differently when language becomes calculation.

At a time when AI systems sit at the heart of ecological, geopolitical, and informational upheaval, this lecture makes the case that the humanistic analysis of machine meaning-making is not ancillary to our present crises but able to be playing a central part in it and through it.

All are warmly welcome to attend.

We will be live-streaming the lecture. If you would like to join us remotely, please choose "Live Streaming Ticket" at check out.

Abstract

Language generated by ChatGPT and other large language models is not simply human language reproduced by machines: it is something more consequential. Drawing on examples from contemporary AI safety research—including safety practices and interpretability work from leading AI labs—Bunz examines why aligning AI systems with human values remains a task difficult to resolve. At the same time, efforts to control harmful outputs reveal something deeper about calculated language and our future with it.

If the humanities in the twentieth century were concerned with ‘symbolic culture’, their task in the twenty-first is to understand ‘vector culture’. What does it mean when meaning emerges from AI systems and their calculations? What, precisely, is changing? While AI systems sit at the centre of the ecological, geopolitical, and informational crises that define our present, this lecture suggests that analysing their meaning-making and understanding their internal logic may also help us to imagine new ways forward.