In this episode, we'll explore the evolution of the essay as an assessment tool and ask whether AI marks its demise or its transformation. The AIcademia podcast is a weekly show helping educators like you leverage AI in your everyday practice. I'm your host, Andy Fisher, and thanks for joining me.

Today’s topic was inspired by a conversation I took part in on LinkedIn, following a post by Paul Matthews. He had an image of a hand holding a sheet of paper that said ‘don't cancel the essay’ and in the body of the post, he invited a comparison between the following two statements;:  statement number 1 - ‘AI can write essays, therefore we shouldn't teach essay writing’ and statement number 2 - ‘machines can lift more than humans, therefore we should cancel PE’. The two are comparable of course and highlight the fallacy in such thinking. His point being that developing and practising skills are not irrelevant just because they can be duplicated by artificial intelligence.

Mathematics, spelling, coding and memorisation are all useful competencies because the focus is not only in the outcome. Paul comments that he gets his students to write essays not as products but in order ‘to produce a certain kind of person’. In other words, he is interested in the process of essay writing which is where he perceives the real educational value to lie because the long form essay affords a learner the opportunity to think in ink.

The conversation inspired me to take a look at the history of the essay. It really came into its own as a writing form with Michel du Montaigne, a 16th century French philosopher and writer. He produced a series of essai, which is the French noun for an ‘attempt’ or ‘effort’. He began writing these personalised reflections in the 1570s and continued until his death in 1592. They were not definitive treaties, so much as meditations where he tested his ideas on the page without necessarily arriving at any particular conclusion. He wrote in an effort to better understand himself and his relationship with the wider world.

Now there were precursors like Plutarch, Seneca, and Augustine, who wrote digestible chunks of philosophical material, but it was Montaigne who personalised the essay, made himself the subject of inquiry rather than moralising or adopting a didactic tone. His writings were conversational, unpolished, secular and open ended. Montaigne was writing at a time of great tumult with religious wars being waged between the Catholics and Protestants, he retreated to his estate in Bordeaux following a successful career as a magistrate and dedicated himself to addressing big questions while avoiding dogma, which would have been dangerous at the time. He identified that his writings became a kind of personal therapy where he explored his own inner life.

109 essays in total survive and reflect his humanist philosophy and forward thinking. Examples include ‘Of Cannibals’, a critique of the European arrogance towards so-called savage cultures, ‘Of Friendships’, ‘Of Smells’, ‘Of Thumbs’, and ‘Of the Education of Children’, in which he recommends of experiential learning rather than learning by rote. The shift from Montaigne's writing to the form that we now recognise as the academic essay still had centuries of evolution to follow. Writers like Francis Bacon, Voltaire and Emerson added their voice to the form as we moved through the later Enlightenment period, but it only really emerged in its final form in the 19th century with the foundation of modern universities and the birth of the liberal education.

These essays were sometimes referred to as ‘compositions’ or personal arguments and afforded an opportunity for students to demonstrate their appreciation of a topic or argue their point of view on a subject given by their teacher. Today's five-paragraph essay, with its standardised rubric and assessment objectives, is a distant cousin to Montaigne's freewheeling and more introspective style.

But now, with AI capable of generating seemingly perfect essays at the click of a button, many educators are asking: does this technology spell the death of the essay form? AI tools like ChatGPT can produce essays that are grammatically flawless and contextually sound and this has understandably sparked concern about academic integrity. Some universities have even taken hard-line stances, ready to expel students caught using AI in their assignments, irrespective of whether the entire submission was AI generated or if AI tools were used to augment the candidate’s work, assisting in the editing and proofreading process for example.

I read a post recently by a university professor bemoaning the fact that she spends inordinate amounts of time reading theses with questionable provenance but feels powerless to award merit where it is due and to penalise those who are cheating because, as plenty of research has shown, AI detectors are not fit for purpose. There are plenty of laughable examples of students accused of plagiarism when in fact the detector merely recognised that English wasn't their first language or they had a special educational need which meant that their writing style didn’t conform to patterns of standard English. Even the Declaration of Independence has been identified as 98% AI generated by some detectors.

If you haven’t yet read Richard Culatta’s excellent article published in ascd in February 2025, entitled ‘The Real Way to Stop Cheating in an AI World’, I’ll make sure to include a link in the show notes. He states, and I quote, ‘the sudden alarm to do something about a problem we've been complacent about for decades highlights some fundamental misconceptions about both why cheating happens and the role technology plays or doesn't in enabling it’. 

Richard goes on to say that cheating is not caused by technology, it's caused by culture, and this is an idea which I find myself very much aligned with. He says that there are five primary reasons why students cheat. 
The first is that the assessment feels irrelevant or lacks real world value. 
The second - academic integrity and best practises haven’t been taught clearly enough. 
Third, there's been too much of a focus on grades rather than learning. 
Fourth, there's high pressure and anxiety around results achieved. 
And fifth, students feel no agency in the assessment process, and he addresses each of these in turn.

He says in order to address the issue of irrelevance and value, we should design more meaningful assessments that students can recognise are aligned with real world problem solving or challenges. 

In terms of the lack of academic integrity, he says that this must become normalised with student input, so there should be negotiated contracts of what constitutes plagiarism. I think the Ship of Theseus paradox is worth considering here. This analogy asks if Theseus's ship was repaired so that one piece was replaced at a time until none of the original ship remained, does it remain Theseus's ship and if not, at what point did it become something else? 

This conundrum carries over to the question of what we think constitutes plagiarism. If a student uses outside sources for research and incorporates those in their own language with appropriate attribution, is that cheating? If they use a spell checker or a proofreading tool built into their word processor, is this cheating? If they engage in peer planning and review, reading one another's drafts and giving verbal feedback so that they can improve, is this cheating? And if AI were to fulfil each of these functions, is it any more or less congruent a process?

Writing an essay involves a number of steps which I'll return to later in this podcast, and part of academic integrity is understanding what is or is not acceptable practice in our institutions, subject,  and course . If we don’t have clarity as educators, we can hardly blame our students if they misinterpret acceptable best practice.

In addressing the excessive emphasis placed on grades rather than learning as a driving factor for cheating, this is really about the philosophical approach that we have in our classroom. Are we teaching to the test or are we using assessments only as a metric of the learning which is at the heart of our curriculum? Of course, this is a bigger question that I'd like to return to in a future episode when we look at whether it's time for us to revisit what is taught and how we teach in schools. 

The motivation to cheat due to the anxiety and pressure to attain high grades is also a thorny challenge. Perhaps, Richard Culatta suggests, we should focus more on low stakes assessments with an emphasis on mastery of learning.

I have some sympathy for this idea, but I also recognise that there's a difference between formative assessments set and marked by a teacher, and the summative assessments that we're sometimes preparing our learners for, whether they are coursework submissions that weigh significantly in their overall exam performance, or a thesis submitted at the end of tertiary education. It's harder to argue that we can eradicate anxiety and lower the stakes in these cases. Instead, as I will go on to argue, the issue may be more to do with developing resilience and managing expectations in our learners while also ensuring that we as teachers don’t exacerbate the obsession with grades.

Finally, in addressing ways that we can help students feel more agency, Richard says we should involve them more in the design and assessment process. We should give them options for assignment topics, or the way in which they would like to demonstrate their learning and understanding.

Ruminating on  Paul Matthew’s post, Richard Culcatta’s article and some of the other reading I’ve been carrying out this week, I’m confident that we are not hearing the death rattle of Montaigne’s beloved essay. In fact, AI might actually be inviting us to return to the essay's original purpose - as a means of personal reflection and genuine intellectual discourse.

And with this in mind, I’d like to propose four practical ways we can adapt the essay writing process to take into account technological advances so that it remains relevant in schools today. I’ll unpack each in turn but here’s the cliff notes:

1 We should consider creating plagiarism-resistant tasks with real world relevance and learner buy in
2 We should adapt our rubrics to reward originality and personal engagement over mere information retrieval
3 We should revisit the conditions under which essays are researched, planned and produced
4 We should explicitly teach and model an ethos of academic integrity

So let’s start with the issue of task design which is plagiarism resistant. Whenever possible we should provide a range of tasks or invite our learners to compose their own titles which resonate with their areas of interest. 

We could ask students to ensure those titles focus upon current events or specific case studies which have been used in the lessons preceding the writing task, or we could refer them to texts and resources which are paywall restricted but which have been made available to our students, because these are less likely to be accessible to large language models and will not be built into their data sets. 

We can also ensure that students personalise their content, drawing upon their own experiences in their writing, and if we know our students well enough, we will recognise these authentic elements. Submissions like these would also be quite refreshing to read rather than ploughing through reams of generic regurgitations of our own class notes fed back to us which might as well be AI generated for all the original thought or personal engagement they reflect.

Here are a couple of example essay titles which demonstrate these principles:

‘With reference to one of the role plays we conducted and two of the speeches we explored in our recent lessons, explain the effect of rhetorical appeals in Shakespeare’s ‘Julius Caesar’.

Here’s another - ‘Express your views on the principle of Universal Basic Income. In your answer, make direct reference to some of the arguments presented by your peers in our class discussion and draw upon the 3 sources linked on the School’s Intranet’

And one more - ‘Drawing upon the work we have done on coastal erosion, choose one of the local beaches in our county and explore the various ways in which humans have impacted that specific environment in the last century.’

Moving on, we should consider adapting the rubrics that we use so that they award originality, personal engagement and critical thinking. For example, we can require a unique perspective, or we could have a self-reflective component of the essay where the learner critiques their own argument or articulates its origins, and that could be built into the mark scheme. Again, for formative assessment tasks these rubrics could be crafted in collaboration with learners rather than being imposed by the teacher to improve the chances that they will be invested in the assessment process.

We can also consider ways in which we can adapt the conditions under which essays are written. 
One well-practiced approach is that we demand controlled writing conditions, so that pupils write long-form handwritten responses in the classroom or on devices which are not web-enabled. We could create lessons where pairs or small groups collaborate on the planning and research elements of the process before then writing their draft independently. This interim step overcomes the tyranny of the blank page which can lead some to take the shortcut of using AI indiscriminately. 

We can also arrange conditions so that there is more of an emphasis placed on teachers having oversight of version control and we could ask for staged submissions, including the handing in of a proposed outline with intended sources, and then the draft could be handed in with marginal notes to show how their argument has evolved over time.  Finally, we could include a viva or some other kind of post-submission discussion where students are invited to defend the ideas that they've presented in their essays. 

I should say that none of these first three interventions will guarantee originality of content. As technology continues to advance, shrewd learners who are invested in bypassing these kinds of checks will find opportunities to do so. We can already deploy agents that will transcribe in real time so that keystroke version control can be bypassed. If Elon Musk gets his way and a significant portion of our students are augmented with neuralink implants or wear smart glasses, it will be even harder to discern the provenance of any idea articulated by a learner in real time. The point here is not to eradicate cheating but to reduce the motivation to do so which leads me to my last, and perhaps most important thought on this topic. 

I think as educators we must engender a culture of academic integrity where we make it very clear what our expectations are and how those expectations can be demonstrated by good practice and digital literacy. We can discuss what the purpose of the essay is, so that students are more likely to buy into the value of this form of assessment, and we can model good practice by planning, writing, and sharing our own essays with our students. 

In the build up to the GCSE study leave for Year 11 pupils here in the UK, I will often write under timed conditions along with my class and then photocopy the output, warts and all for my students to read. It keeps me on my game, hopefully shows how to write to the demands of the mark scheme and it engenders respect from those pupils who need to see a teacher walk their talk before they will trust what they have to share. As technology advances it will be harder to prevent cheating - our most effective intervention therefore is to create a respect for learning and an culture of integrity so they don’t want to.

So, to wrap things up today, I don’t think AI will bring about the death of the essay. I'd argue we're actually seeing its renaissance. The essay is evolving, just as it has since Montaigne's time. The key difference is that now we are invited to return to its roots as a writing form - emphasizing personal insight, critical thinking, and authentic voice over mere information delivery.

Thanks for listening—I hope you've found some useful takeaways from the conversation. Please do spread the word if you think others would like the show, and do check out the AIcademia YouTube channel where you'll find practical tutorials that complement the topics covered on this podcast. Have a great week, and I look forward to catching up again soon.