mobile phone on a marble surface with ChatGPT app open
AI and Education: ChatGPT and its Future

AI and Education: ChatGPT and its Future

By Liam Needham, eLearning Support Officer: Faculty of Biology, Medicine and Health, The University of Manchester

Artificial intelligence has long been touted to change the world as we know it. Long gone is the excitement surrounding chess playing supercomputers; the newest cool kid on the AI block is ChatGPT. But what exactly is this fancy new tool that everyone is talking about, and more importantly, how might it impact higher education?

ChatGPT is a large language model chatbot, developed and released by the company OpenAI in November 2022. The tool is able to generate realistic, human like text based on any prompt provided by the human user. For example, ask it to explain how bone formation occurs in humans, or how to carry out an ELISA analysis, and you’ll get detailed, well written and well articulated responses. Ask the model to adapt those answers for an audience of five year olds, and you’ll get vastly simplified, but accurate descriptions of those complex processes.

It’s this extraordinary ability to articulate detailed responses that has the education sector talking. There are fears in some circles that students of all ages will simply lean on their new AI friend for answers when asked to complete essays, coursework and even remote examinations. This is particularly feasible given that at the time of writing, anyone with access to the internet can freely access ChatGPT and its wealth of knowledge.

Various studies in the sector have already demonstrated that the tool is capable of passing exams in various disciplines, including law exams (Choi et al., 2023), multiple choice macroeconomics exams (Geerling et al., 2023) and even the United States Medical Licensing Exam (Kung et al., 2022). ChatGPT is not passing these exams with flying colours. But in the world of examinations, a pass is a pass.

a pile of harback books from the point of view of their pages only
(Dato-on, 2023)

Despite this, it’s relatively easy to spot some of the areas the virtual genius lacks:

  • Lack of referencing – ChatGPT does not divulge where exactly its information has come from, given that it effectively amalgamates it from an infinite amount of sources. This would leave students, for example, in a position where they have to find suitable sources themselves, assuming that they want to achieve a high mark in your typical higher education assessment.
  • Time limited knowledge – The tool was only ‘trained’ on information up to September 2021, meaning it cannot comment on any events or facts that have occurred after that date. For example, if you ask it to explain why Liz Truss resigned as Prime Minister of the United Kingdom (in October 2022), you’ll be told that she has never been Prime Minister of the United Kingdom.
  • Lack of internet access – It may not be entirely obvious to new users, but ChatGPT itself cannot access the internet, meaning real time news, websites, location based information and much more remain inaccessible. The model simply pulls its knowledge from the material, as per OpenAI’s training (e.g. books, websites).

Given these issues, the collective mindset across the sector has generally turned towards questions such as: ‘how do we stop students using ChatGPT to cheat on their assessments’; ‘ChatGPT could actually prove a very useful tool?’; ‘How can both staff and students utilise ChatGPT to enhance their teaching and learning?’.

At the simplest level, students may utilise ChatGPT to give them an initial overview of a new topic, or for simplifying complex ideas and processes they may not fully understand. It could even be used as a virtual assistant of sorts, guiding students through ways in which they can improve their writing, or in the context of computer programming, suggest improvements to the code provided by the user (yes, it can do that too!).

The benefits of ChatGPT aren’t restricted to student users. At the mere touch of a button, the chatbot can generate a full lesson plan based on a prompt, including timings, suggested materials and lesson objectives. Critical feedback on student work is yet another string to ChatGPT’s bow, with a particular emphasis on grammatical and structural advice. It does come with its own warning however, stating that the feedback provided may not be as comprehensive as that of a human.

a partial image of a person in a graduation gown, holding out their graduation cap and scroll
(Danilyuk, 2021)

So, where does this leave the education sector? The truth is, there’s no definitive answer just yet. You only need to head to your favourite search engine, put in the words ‘ChatGPT’ and ‘education’, and you’ll be greeted with a myriad of articles debating how education might have to change, how assessments will adapt, and how students may actually just breeze their way through three years of an undergraduate degree (however unlikely).

Here at the University of Manchester, colleagues in the Faculty of Biology, Medicine and Health are leading a study to explore the perception of AI in Higher Education assessment, with plans in place to trial artificial intelligence tools across various taught courses to gauge their usefulness. Research of this type will undoubtedly be firing up across a whole host of higher education institutions, as we are all keen to find the ways in which we can make AI work.

So, while we wait to see all of the ways in which ChatGPT might shift the higher education landscape, I’ll encourage you to sign up and give it a try yourself. I think you’ll find yourself pretty amazed at what it can do for you.


Choi, J.H., Hickman, K.E., Monahan, A and Schwarcz, D.B. (2023) ChatGPT Goes to Law School. Minnesota Legal Studies Research Paper. SSRN: January 23, 2023; 23-03. Available at: (Accessed 23/03/2023)

Geerling, W., Mateer, G.D., Wooten, J. and Damodaran, N. (2023) ChatGPT has Mastered the Principles of Economics: Now What. SSRN: February 13, 2023. Available at: (Accessed 23/03/2023)

Kung, T.H., Cheatham, M., ChatGPT, Medenilla, A., Sillos, C., De Leon, L., Elepano, C., Madriaga, M., Aggabao, R., Diaz-Candido, G., Maningo, J., and Tseng, V. (2022) Performance of ChatGPT on USMLE: Potential for AI-Assisted Medical Education Using Large Language Models. medRxiv 21 December 2022. Available at: (Accessed 23/03/2023)

Images from Pexels

Did you enjoy reading this? To become a member of our community see Membership details here

Leave a Reply

Your email address will not be published. Required fields are marked *