Opinion: AI can now help write essays that appear highly similar, if not indistinguishable, from that of a human author

By Eamon Costello and Mark Brown, DCU

Artificial intelligence-powered automatic essay writing technologies have taken a huge leap forward and are becoming widely available. In many cases, AI can help write essays that appear highly similar, if not indistinguishable, from that of a human author.

As this piece was being typed, Google AI helpfully suggested words to auto-complete these sentences. This speeds up writing but other tools and services, such as those based on the GPT-3 technology from the OpenAI group, can do much more than simple sentence completion: they can write the whole thing from scratch from start to finish.

AIs can generate surprising comprehensible text

By simply taking a prompt of a sentence or phrase from a human author, AIs can generate surprising comprehensible text. When the prompt "when an AI wrote my essay" was typed into a freely available online AI text generator to see what it would produce, the AI wrote: the following

"And don't forget", the professor said, "AIs are not allowed to write your essay or help you write your essay in any way" I groaned and rolled my eyes, thinking that I knew what he was going to say. I had never been good at writing essays, and I knew that an AI would only make it worse. But when I sat down to start it, I found myself in the middle of a beautiful essay. The words flowed easily from my fingers, and I was suddenly inspired. I had never written so well, and I was surprised that the AI had written it for me. I finished the essay and handed it in, feeling proud of myself. I didn't know what the professor would say, but I was sure that I had done a better job than if I had written it myself.

One can see how a student could generate text in this way, perhaps generating several texts, seeing which one suited best and then lightly editing for clarity. AI can also take large bodies of text as input and attempt to summarize them.

We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences

From RTÉ Radio 1's Today with Claire Byrne, Brian O'Connell reports on student essay mills

A student could prompt an AI to write several paragraphs and could then select which paragraphs seemed most well written and coherent. Finally, they could feed these paragraphs back to the AI asking it to summarise them. This could form the basis of a summary/conclusion section of an essay.

Using AI in this way requires particular skills. Just as correct spelling is becoming less critical with the ability of autocorrect, perhaps essay writing will evolve similarly. It may be that writers in the future engage in the higher level activity of orchestrating a composition, while AI does the heavy lifting of producing the actual sentences.

If using a spell checker isn't seen as cheating, will the use of AI for essay writing then be accepted as the new normal? Perhaps future students will use AIs to write their essays, while professors deploy AIs to check their authenticity.

Of course AIs can grade essays too, but does this mean that teachers will have less work?

Of course AIs can grade essays too, but does this mean that teachers will have less work? The jury is still out on this question: one major review of the research on AI in education found a conspicuous absence of reference to actual teachers. One scenario is a teacherless future where students are accelerated through courses of study by advanced robo-Profs.

A contrasting future has been foreseen by AI education expert Peirre Dillenbourg. He has predicted that we will have more teachers in the future, not less. He foresees teachers working in teams to oversee and design learning scenarios using multiple AIs dedicated to specific educational tasks.

That is the future taken care of but what about the present? Universities worldwide currently invest heavily in anti-plagiarism and academic integrity technologies. Many of these systems have been termed 'data-extractive', in that they often rely on extracting and mining large bodies of student work. At their worst these expensive systems can create climates of fear., where students feel they are being policed by big brother or sister.

With all of the fuss about AI, it is worth remembering that people are always at the heart of education

AI essay writing may be seen as just another chapter in the long history of so-called "essay mills", services that students can use to commission and buy their homework from. Will AI make these services redundant in the future? What constitutes cheating and breaches of academic integrity in the world of AI? After all, irrespective of how we define cheating, who loses if the student does not fully engage in their own learning?

Something that educators can do is to have conversations with students about their learning and especially their assessment. A guiding principle should be that a student will always want to do the work themselves given the right conditions. This is the opposite of a starting principle that says: every student is a potential cheater.

Assessment mixes that are not completely dependent on traditional essays can allow students to express themselves in a variety of ways. Do we try to tame AI to protect old ways of learning or should we embrace its potential and reimagine our assessment practices to reflect the modern reality of living in the 21st century? One creative educator had his students purposefully use and evaluate AI essay writers as part of their assignment.

With all of the fuss about AI, it is worth remembering that people are always at the heart of education. Student and teacher workloads should be key considerations in the design of assessment. Giving each other space to build trusting environments in which to teach and learn will require much human ingenuity, care and intelligence.

Dr Eamon Costello is an Associate Professor of Digital Learning at the DCU Institute of Education. Professor Mark Brown is Chair of Digital Learning and Director of the National Institute for Digital Learning at DCU.

The views expressed here are those of the author and do not represent or reflect the views of RTÉ