Thanks so much to Milan Stürmer for this thought provoking and insightful reflection on generative AI for academics:
However, it might be that these capacities are acquired and maintained through just the kind of reading and writing practices that are in danger of disappearing with the widespread adoption of Large Language Models (LLMs). For those that have acquired advanced levels of literacy and trained their scholarly craft prior to their widespread adoption, the distinction between ‘thinking with’ and ‘substitute for’ might seem much more clear-cut than for those born into the age of LLMs. If and how the practice of ‘thinking with’ can sustain its own condition of possibility is still an open question.
Throughout the book, I find myself agreeing with Carrigan’s (2025) enthusiasm on an abstract level, while remaining consistently unable to engage with conversational agents in an equally meaningful and productive manner. The affordances of GenAI systems just seem much less suited to my own routines and habits. Which confronts me, as a reader, with a conundrum: How far am I willing to change my own practice to better accommodate conversational agents as collaborators? This, unfortunately, is unlikely to remain a question of personal preference. If/as these systems get adopted more widely and the academy accelerates even further, it might no longer be a choice, at least for those without permanent positions.
Mark Carrigan 