Why The Open University will win in the post gen-AI world of education

By Doctor Mark Burgin, BM BCh (oxon) MRCGP

As a doctor who has studied management and learning with the OU, medical ethics and law and written a book on the Gen-AI and the PGCME I might have an insight into Generative AI. The technology has a massive advantage and a great flaw which make it difficult to implement. Only distance learning institutions such as the Open University are well placed to deal with both. My alma mater Oxford University has already failed to develop multimedia learning tools to level of the OU.

The massive advantage of Generative AI is that help students learn about 5 times faster than usual. If a student can stay in the learning zone they make exceptional progress. The technology provides feedback, helps students reduce cognitive load and writes eloquently. With the right prompt engineering Generative AI can collaborate with the student and keep them engaged.

The great flaw in Generative AI is that the work is well written and plausible but mostly rubbish. Copying and pasting an answer to an essay question will get the student a passing mark but do nothing to help them learn about the subject. Just like their picture equivalents they are superficial and lack any real substance. There is nothing real within the text so no deeper understanding can be achieved.

The challenge for universities is to learn to harness the potential of generative AI but at the same time avoid the pitfalls. Some universities are locked into old fashioned methods of learning and will struggle to adapt. Most are still having their ‘calculator moment’ where they are having to learn from their students. Deans will need to move quickly if they do not want to face a dramatic fall in their intakes as students vote with their feet.

How to use Generative AI to learn

There is a truth at the heart of every learning experience that it is the connection between the teacher and the pupil that matters. Generative AI can collaborate with the pupil in a similar way to a teacher but the pupil must know how to use it properly. I was asked whether I wanted to do mathematics at my interview Oxford University. I replied that I thought that brains are more interesting than computers.

Generative AI needs the student to have a mathematical understanding before it will collaborate with them. Mathematics is the real language of Generative AI and clever questions will only take a student so far. Only when the student can use prompt engineering can they fully unlock the power of Generative AI. Under the hood Generative AI is powered by large matrices of values which balance the output.

Prompt engineering relies upon the student being able to see the properties that emerge from the calculations. They must identify which part and version of the answer can be further explained. Reconstructing the outputs of the Generative AI and giving just enough new information to get where the student needs to go. The learning is hidden in a maze and many will struggle to find their way.

Whilst the process is second nature to people with a learning style that includes mathematics others will need to be taught. The techniques themselves are not complex but will be unfamiliar to many. It is not enough to ask the student to list the prompts that they have used, they will need help in real time so that they can learn how a good prompt feels. They need to experience that moment of discovery for themselves with support and encouragement.

The Open University

The Open University has always understood their learning tools better than any other institution. The preparation of materials, the way that they are presented and even the questions are all carefully crafted. The Open University’s weakness is that it cannot give the levels of face-to-face education as other universities. It has turned this weakness on its head by providing materials with exceptional insight into the learning process.

The OU will be the first to recognise that unless it teaches its students how to do prompt engineering they will not gain the benefits from generative AI. Whilst other universities try to prevent students from using the technology by plagiarism software and going back to paper and pencil the OU will empower students. Teaching prompt engineering as a mandatory module will ensure that every OU student can boost their learning.

There are many arts students who will find the concept of prompt engineering difficult. The mathematics is off-putting but they have their own way into understanding the technology. Many are natural story tellers and can understand the concepts of prompt engineering as telling a more complete story. Many have a spatial understanding and can see the prompts as shapes and positions.

The OU will also be the first to help the students monitor progress in learning. Currently course work is marked on performance, whether it is well written, has lots of facts, sounds plausible. None of these are associated with deep learning and teachers do not have time to analyse thousands of pieces of student’s work. AI can be asked to provide feedback with a prompt, this prompt can be shared with students.

How to teach prompt engineering

Prompt engineering is a new language and the UK has a very poor record of teaching languages. The usual techniques of providing set phrases to learn or rules will be no more effective in teaching prompt engineering than French. Language learning is best achieved by communication, this might suggest that prompt engineering can be self-taught but the problem with Gen-AI is that it produces rubbish.

Students who use AI to help teach themselves prompt engineering fall into a trap, they feel that they are learning and often produce what appear to be good prompts. The problem is that they are learning rubbish, they are not collaborating with the Gen-AI. They are learning what they think learning from the Gen-AI would be like. The AI gives them the illusion of learning whilst only exploring the areas they already know. Gen-AI cannot teach anything the person does not already know.

This phenomenon where student hallucinates that they are learning is an emergent property of Gen-AI. To address this the teacher can ask the student what new information they put into the model. If they have not given the AI any new facts or concepts then the output will be illusory. In a similar way that it caused AI workers to hallucinate that it was sentient.

The amount of new information that should be added to the prompt is inversely proportional to the student’s current knowledge. Each prompt may take a few minutes to several hours to construct. The more hurried or poorly considered the prompt the more likely it will generate a hallucination in the answer or the mind of the reader. Few people can create more than a dozen real prompts in a day.

Conclusions

The Open University is the only UK university that could create an effective prompt engineering module. Students are having to learn how create prompts themselves and are at risk of plagiarism and hallucinating that they are learning. Developing separate courses for different learning styles is essential to ensure that science based students do not gain an unfair advantage.

Unlocking the potential of a 5 times faster learning will change the landscape of further education in the UK. Courses such as medicine which take 5 years to complete are already showing a gap in performance between those who can and cannot prompt engineer. Many institutions which have failed to take advantage of the technology have found that student numbers are falling, perhaps due to moving to courses which embrace the advantages.

Students will struggle with the technique until it is something that they have grown up with. For most over 12s this means that they need help if they are going to get the advantages of the new technology. Prompt engineering is at its heart seeing patterns so can be taught to anyone who can see a pattern. The students will require the connection with a teacher to make this leap.

The Post Gen-AI World of Education will eventually have Gen-AI as a normal part of the learning process. Students will have shortened courses or have more content and the examinations will focus on understanding rather than performance. Whether UK further education survives and whether the next 6 years of students will gain the benefits is less clear. The Open University is uniquely positioned to lead the way and transform the future of education in the UK.

Dr Mark Burgin graduated from Oxford University in 1987 and studied with the Open University on two occasions in the 1990s. He has also studied for the CPE (law), Medical Ethics, learned Portuguese by living in Brazil. He has written many articles and written books on Personal Injury and the LLMS (your PGCME) and is about to publish a book on Disability. His next book is on psychological techniques.

September 2024

Would you like to contribute an article towards our Professional Knowledge Bank? Find out more.