In the ever-evolving landscape of education, technology continues to reshape the way we teach and learn. Chat GPT, an AI-powered large language model is one of several relatively newly-released EdTech tools that shows clear capability to assist and enhance classroom experiences. With its ability to understand and generate human-like text, Chat GPT is set to transform the way teachers engage with their students. In fact, it has already. In this series of blog posts, we will explore five ways in which you can use Chat GPT to revolutionise your classroom.

In these blog posts, I will include screenshots of Chat GPT responses as well as usable prompts for teachers. Keep in mind, as Chat GPT is continuously updated, responses to these prompts may vary in content or format. Also, the version of Chat GPT used in these blog posts will be the free version, not GPT-4, in order to make these methods more accessible to all teachers.

  1. Personalised Learning Experiences

Chat GPT offers teachers the ability to provide personalised learning experiences tailored to individual students. By using natural language processing, Chat GPT can analyse students’ responses, identify their strengths and weaknesses, and adapt instructional materials accordingly. For example, teachers can assign interactive exercises through Chat GPT that adapt based on students’ performance, providing targeted practice and immediate feedback. This personalised approach fosters better engagement, comprehension, and academic growth among students.

Above is Chat GPT’s description of its own capabilities, so I decided to put them to the test by asking it to “give an example of what teachers could input” to demonstrate a personalised learning experience. Here was the result:

As users will find with most prompts given to Chat GPT – the more specific, the better. Without being given specific examples of student work here, Chat GPT can only emulate what it might say if prompted to do so, but the teacher would need to give it specific prompts – such as analysis, then generating a report, etc. – in order for it to complete the tasks laid out above. I trialled this, and despite a few hiccups, the results were impressive.

My initial input was: “Chat GPT, I am going to give you 8 student written assignments for an EdTech class. The students were assigned a 150-word minimum response to topics that were brought up in class. Their responses needed to pertain to EdTech, but I wanted them to prioritise reflection and deep thinking about the topics, not just summary. Can you analyse the written responses, identify their strengths and weaknesses, and put the results of your analysis in a neatly organised table?”

Strangely, I received a table with strengths and weaknesses presumably inferred from made-up content immediately. With a quick downvote (clicking the thumbs-down icon) and explanation, Chat GPT fixed its answer, as you can see below.

I proceeded to enter sample student responses. Below is a section of the analysis I received in response. It is worth noting that this approach will require student responses to be in text form already (difficult in some subjects) in order to make copy-pasting into Chat GPT feasible and timely. Of course, if the teacher has already done the work of analysing student strengths and weaknesses, then they can simply ask Chat GPT to generate further examples for practice in those areas of weakness.

As you can see, the analysis is quite detailed, and clearly shows engagement with each work. The most remarkable thing about the analysis was the speed – it could never be matched by a human, which is what makes it such a valuable time-saving tool for teachers. That is not to say it’s perfect, though. Some of the analysis is rather generic, and I didn’t feel like it considered what I told it while introducing the topic – namely that I wanted to prioritise reflection and deep thinking.

As if to confirm my suspicion, a message at the bottom of the table read: “Please note that the strengths and weaknesses are general observations and may vary depending on the specific criteria provided for the assignment. If you have any specific criteria or guidelines you would like to add to the analysis, I can include them in the table.”

So I followed up with: “The only guidelines I had were that their responses needed to pertain to EdTech, but I wanted them to prioritise reflection and deep thinking about the topics, not just summary. Can you add that into the analysis? Or was it already included?”

To which it responded with an apology and a new analysis:

Great that Chat GPT incorporated the guidelines after a reminder, but I don’t know why it didn’t do that from the outset. This proves the point that you can’t be lax with your instructions – make sure to be very specific about what should be included in Chat GPTs output, and then double-check it to ensure it really is there.

A further issue I noticed is that with the new guidelines, the same weakness came up in all 8 students’ ‘weaknesses’: “Limited reflection or deep thinking beyond summarizing information.” Although I had come to the same conclusion about the majority of the responses, I picked a variety of student submissions, some of which engaged in reflection and deep thinking much more than others. One hypothesis I have about this result is that the responses were too short to warrant what Chat GPT considered reflection/deep thinking. Alternatively, it may have been because I didn’t define reflection/deep thinking in the first place, but needing to specify such parameters is an inconvenience.

Finally, we get to the personalised learning experiences part. I prompted: “Overall, what might be some areas of improvement for these students, and what type of exercises/practice should I do with them as their teacher?” Below is a section of the response.

The time saved with this type of analysis almost makes it possible to differentiate assignments for each pupil, but that would be a mistake! As we discussed in our Explained series video on Adaptive Teaching, differentiation can lead to varying outcomes for children, widening the gaps between disadvantaged students and their non-disadvantaged peers.

What can be done, however, is to use this analysis to give pupils feedback, or encourage them to independently follow the suggested exercises/practices in or outside of class for a defined period of time. For time-poor teachers (which often seems like all of us), I prompted Chat GPT further: “Since I don’t have time to tutor these students one-on-one, can you provide some strategies that I might use to target their weaknesses all together?”

Though these suggestions are helpful, and Chat GPT can no doubt elaborate on any of them, they did feel a little bit generic, as if ‘EdTech’ had just been inserted into a template of classroom activities. Nevertheless, being able to adapt Chat GPT output to teachers’ and students’ needs on the fly is incredibly useful.

Your turn – try out Chat GPT now

Now you should have an idea of how to prompt Chat GPT for personalised learning experiences. Be specific with prompting, but also don’t be afraid to be conversational – it is designed for natural language interfacing. In other words, use Chat GPT like a tool, but treat it like a person. Remember, never input any personal information into Chat GPT, but do familiarise yourself with its capabilities. The more you play around with it, the more you will realise the time-saving and educational potential of this powerful generative LLM (large language model) AI.

Any questions? Comments? Suggestions? Feel free to send me an email me at

About the Author:

Luke Kemper

Luke Kemper is Insight and Intelligence Lead at HEP. He recently graduated from the University of Cambridge with an MPhil in Education, Globalisation and International Development. Before that, he worked for seven years as a university lecturer and high school teacher in China and Poland.

HEP Talks Podcast

The voice of Haringey Education Partnership. A weekly briefing on the latest stories in education news, deep diving into developments and interviews with leading voices in Education.

More Insights