Unless my friends are particularly fiendish, I’m pretty certain that “I’m just going to ChatGPT this essay” is a phrase we have all heard. I would like to posit that it is no wonder many students succumb to this temptation when the work we are required to do is often as pointless as Elon Musk’s attempts to get people to start calling Twitter ‘X’.
It is not outlandish to claim that in the academic realm, final exam results are virtually all that most students prioritise. It is also not outlandish to claim that our peers are somewhat sensible for doing so. After all, we exist within an education system in which success is largely predicated on the maximisation of percentages. Thus, if an individual is tasked with producing sixteen essays over eight weeks, half of which will cover topics not on their final examination, the rationale behind focusing on non-examined topics becomes unclear.
Of course, an ideal world would see us all so enthralled by our degrees that we want to produce a beautiful piece of writing on every topic, purely for the enjoyment of learning and the development of our writing skills. In some sense, many people I know do complete their work with this intention. An issue arises, however, when this intention clashes with the desire to lead a somewhat well-rounded life, and also to get enough sleep. Call me impulsive, but I truly believe that when faced with the decision to enjoy a unique evening with your friends or to dedicate yourself to a dull essay topic that is not going to contribute to your final grade whatsoever, it is much more sensible to enjoy the evening with your friends.
I would like to make it clear that I am not advocating for such scenarios to be actively sought out. I believe that most students possess the necessary time management skills for the use of large language models (barring the use of them as a thesaurus, spell-checking device, or Google-substitute) not to be a common, necessary occurrence. I do however believe that a key reason LLMs are used is because we already fulfil our personal learning goals by engaging with examined topics, and therefore superfluous topics are logically relegated below a good night’s sleep.
I admit I am speaking from a rather Oxonian lens. There are likely students at other universities–at which it is only required that students produce two or three essays per term–who still use ChatGPT as their heavy-handed assistant. Such cases do draw out my inner traditionalist.
Of course, there is also a fine line between optimisation and over-optimisation. In machine learning, “overfitting” occurs when an analysis is too tailored to a specific (training) dataset, preventing it from accurately generalising to new data. Similarly, we must be cautious not to overfit ourselves when it comes to prioritising optimisation – optimising our lives to the extent that we lose the ability for creative thought and production. If we do not practise thinking for ourselves, I believe we may begin to lose ourselves. A loss of grit is not a trivial matter, and the sense of satisfaction obtained from completing a piece of work as an individual cannot be replaced. If you have found yourself succumbing to the pull of idle half-participation in your work, I thus challenge you to do your next piece of work without using an LLM. You will very quickly feel the positive consequences on your ability to think and communicate clearly. I also genuinely believe it will lead to you feeling much happier in the long run as you become more engaged with your own existence.
Fundamentally, I believe most people would prefer to complete their work without the use of an LLM; the use of them for academic matters is typically a symptom of stress. To tackle this stress, course directors could, for example, trial the removal of written work on non-examined topics. Nonetheless: whatever the future of education holds, antagonising students for using ChatGPT is not the solution. To preserve creativity, we should address the root causes of students’ reliance on AI while encouraging original, human thought.