I’m long past the creative exuberance which I felt over the summer as the bulk of Generative AI for Academics poured out of me. It was a strange experience to have a full book pop into my head fully formed, as opposed to being something I discover through the act of writing. I’m now deep into the most agonising editing process I’ve ever experienced because (quelle surprise) if you write 90,000 words in the space of a few months, a significant portion of them are crap and the text as a whole doesn’t fit together. But I’m rapidly getting there and I’m starting to feel I could be done within the month.
I was getting Claude’s help with finalising the first two chapters. I asked it to list the principles I had proposed in these with a view to making sure I thread them consistently through the book as a whole:
- Adopt a reflective approach to incorporating generative AI into scholarship rather than passive or uncritical use
- Treat conversational agents as interlocutors and collaborators to think with rather than as passive tools
- Focus on using generative AI for enriching thinking, communication, collaboration, and engagement
- Develop prompts carefully to guide conversational agents productively rather than relying on one-off instructions
- Build ongoing dialogues with AI systems to establish shared understanding and goals
- Experiment iteratively with different AI tools and methods to find what works for your goals and workflow
- Consider both the creative opportunities and risks/challenges presented by generative AI
- Focus on process over product – integrate AI into scholarly workflows rather than just evaluate outputs
- Develop technological reflexivity rather than taking infrastructure for granted or viewing AI as an add-on
- Be strategic in tracking AI developments rather than seeking to comprehensively “keep up”
- Prioritize developing satisfying and sustainable AI uses that align with your values and priorities
- Explore how generative AI can enhance scholarly activities like note-taking, drafting, collaboration, etc.
- Consider the broader economic/political/social context and impact of mainstreaming generative AI
It occurred to me that condensing these down to 5 points could make for a useful presentation: 5 principles for generative AI for academics in which I spend five minutes on each one. But basically I want to be done now because it’s impossible to get real editing done during the week in term time, which means I have little weekend to speak of until this book is finished.
