Large language models (often wrongly referred to as “(generative) artificial intelligence”) are mindless algorithms, trained on the output of countless human language users, and programmed to regurgitate statistical averages across this training data.

LLMs do not observe, they do not think or feel. They do not know, they do not understand, they do not reason. They cannot be right or wrong. They do not create. They replace real human thought and speech with a travesty of human thought and speech.

LLMs serve no-one except for the technocrats that deploy them. Their use is antithetical to everything the arts the humanities and the sciences stand for — the values of enlightenment, the creativity of the human mind, the engagement with human cultures. They hinder the development of humanity as a whole and the individuals that use them. They have no place in teaching, research or art.

In work submitted in our lectures and seminars — contributions to in-class discussions, presentations, essays, term papers, etc. —, large-language models may not be used for any purpose. This prohibition covers not only the use of LLMs as an alternative to human brainstorming and research or their use to generate text that is subsequently used as part of written or spoken assignments, it also extends to the use of LLMs for the purpose of translation, summarization, proof-reading and other seemingly low-level cognitive and linguistic tasks. If we have reason to suspect that you have used an LLM in any of these or other ways, we will remove you from the lecture or seminar.

It is our job to aid you in developing your capacity to think, to read and to research, to understand and to argue, and to write. Our aim is to help you achieve your potential as genuine contributors to human culture. This aim is made all the more urgent by the very spread of LLMs through all domains of human culture — we need to help you develop these skills so that you can use them long after the LLMs have poisoned the libraries of human knowledge and art, long after they have reduced their users to mindless and helpless prompt-givers.

But you have to want it too.