Welcome back! In earlier parts of this season, I’ve argued that gen-AI is really just an immensely sophisticated word-prediction machine. I’ve demonstrated how inappropriate use has led to severe consequences for professionals, and I’ve considered the implications for students. The last post argued that the struggle you experience to understand, analyse and evaluate is the very essence of study and not to be bypassed. It’s what university study is for.
In this post I want to explore what happens to us as individuals when we outsource our thinking to the machine.
How we learn
I made a strong claim in the previous post: the struggle is the essence of study. Struggle can show up in different ways during your undergraduate law studies and it is legitimate to ask whether a struggle can be avoided. I’m here to argue that some of those struggles should, in fact, be embraced.
Reading to think
You will always have more to read than you can read. This is a key struggle faced by students. There simply is no way, in the time available to most people, to read everything on the reading list in its entirety. Yet there is no understanding, analysis or evaluation without doing the heavy lifting of reading. Students often imagine that reading sits outside the real work of studying. It does not. Reading is where studying begins.
Seasons 1 and 2 of this blog provide you with strategies and techniques to manage the information overwhelm. Key takeaway: you don’t have to read everything word-for-word. Have a reading strategy, know what you are looking for, and treat your sources like an annotated dictionary. You wouldn’t dream of reading the dictionary word-for-word. Don’t do that with the textbook, cases and articles. Instead, look for key principles, arguments and reasoning. Seasons 1 and 2 guide you how to do that.
Reading supplies the raw material for thought. Writing is where that thought becomes visible.
Thinking to write/writing to think
True thinking requires voice and you need to find yours and gain confidence in using it. Your thinking voice shows up in (at least) three ways:
Thinking to yourself inside your own head requires listening to your internal voice and your ability to focus deeply on a problem. This is uniquely challenging in a world of easy distraction but it is an essential part of working out what you believe about an issue.
Thinking out loud with others requires you to formulate your thoughts in real time, while integrating what other participants in the conversation contribute, so that you can speak meaningfully when it is your turn. This is uniquely cognitively demanding because of the multiplexing of your thoughts with those of others. It is also very helpful for developing your thinking, because the meeting and clashing of ideas sharpens your arguments.
But your most consequential thinking will be done on paper or on screen, when you turn your thinking into a permanent record crafted for others to read. Sometimes the words will come out in a stream of consciousness; the fabled writers’ flow-state. Reaching this state is so much more likely if you have immersed yourself in the literature and wrestled mentally and orally with the arguments. But more often than not, you will need to ponder, play around with a thought to get the right word, delete, reframe, write messy-drafts, iterate, improve and keep doing it until the thought is fully developed and made manifest on the page.
This is truly hard and everyone finds it a struggle, at least sometimes.
No wonder it is tempting to dump the problem into a gen-AI chatbot and outsource the hard graft of truly thinking things through. The AI is so plausible, so smooth and so downright easy that once you see its suggestions the temptation not to do the thinking yourself can become overwhelming.
So what does dumping the problem into gen-AI do to us?
I want to suggest that outsourcing your reading, your writing, your thinking about a topic, diminishes something much more important than any of those things in isolation. It diminishes your agency. We lose something profound when we outsource reading, writing and the thinking that go with them: the joy of discovery; the lightbulb moment; the challenge of a new argument; the ability to hold a meaningful conversation about a serious topic, carry an argument or convince others of a truth. These are the things that develop our agency and make us interesting people! Don’t discount these much more personal consequences and the emptiness of being on the outside of conversations where others seem so effortlessly informed and erudite. The effort may be invisible, but it is not absent.
The student who relies on gen-AI to produce explanation, analysis and evaluation is not performing those things themselves and ends up circumventing the development of their agency and eventually diminishes it. Now that is a philosophical argument, but there is already some tentative empirical research from MIT supporting these concerns. The study found LLM users had weaker neural connectivity, a lower sense of ownership of essays, and struggled to quote from their own work.[i]
While this does not demonstrate that using gen-AI diminishes your agency per se, the findings are certainly consistent with the broader concerns I am raising, in this article specifically, and the series generally. Something important may be lost when we outsource our learning and the work involved in constructing and articulating our own ideas.
So, let’s remind ourselves the arguments of each of this series of articles:
- Don’t outsource your thinking to the machine.
- Don’t outsource your professional reputation to the machine.
- Don’t outsource your intellectual development to the machine.
And finally:
Don’t outsource your agency to the machine.
Which leaves just one thing remaining.
How on earth do you harness the obvious power of gen-AI without outsourcing your agency, intellectual development, professional reputation and ability to think?
Subscribe to the blog to get notified of the final instalment in this series, where I will address that question head-on.
[i] Nataliya Kosmyna, Eugene Hauptmann, Ye Tong Yuan, Jessica Situ, Xian-Hao Liao, Ashly Vivian Beresnitzky, Iris Braunstein, and Pattie Maes, ‘Your brain on chatgpt: Accumulation of cognitive debt when using an ai assistant for essay writing task.’ arXiv preprint arXiv:2506.08872 (2025). (Read abstract)
Enjoyed this article?
Read Law Right is a blog about reading effectively, writing persuasively, and developing the understanding, analysis and evaluation skills that help law students to succeed.
Subscribe below to receive notifications of future posts.
Leave a Reply