How AI can transform computing education

How can teachers harness the power of AI to improve computing education? 

Prof Miles Berry shared inspiring ideas and insight at CAS’ recent event in London.

The Transforming Tech event held in June included a presentation on AI in Computing Education from Prof Miles Berry, Professor of Computing Education at University of Roehampton. 

He looked at the transformative potential of AI in developing children’s problem-solving and creativity and how it can be used to teach coding and used within established teaching resources such as Scratch.

Miles began by referring to the legal requirement that the school curriculum prepare pupils for the opportunities, responsibilities, and experiences of later life, and so now needs to find ways of addressing the emergence of AI, including through the computing curriculum.

Miles discussed the growing redundancy of the widely taught model that describes computers as machines that take an input, run it through a programme, and produce an output, and suggested that this now needs to reflect how input and output are related through a machine learning model derived from training data. He quoted Matt Welsh’s CACM article from January, which suggested that most software will be replaced by trained AI systems.

“The conventional idea of writing a program might be heading for extinction,” he said.

He discussed what he categorized as ‘old school machine learning’, discussing how models can be trained to recognise patterns and thus decode or categorise speech or images. He provided a demonstration of how this could be achieved through the use of Scratch, and encouraged the idea that children should code their own software using pre-trained models. Miles looked at the emergence of Chat GPT, demonstrating how it can change its tone of voice and write with the appearance of personality and variety. He sketched out the underpinning large language models on which it’s based.

“There is wit, there is humour inside this language model” he observed.

He then went on to talk over the ethics around pupils’ use of Chat GPT.  He stated that most pupils already realised that using Chat GPT to write coursework or essays would be considered cheating but suggested that using Chat GPT to explain something, provide ideas or suggest improvements to work may be more acceptable, or even beneficial.

More guidance is clearly needed in this area, as teachers and assessors are advised to only accept work for assessment which they consider to be the students own. “That’s putting a lot of responsibility on teachers’ shoulders,” said Miles, especially given the difficulties of verifying whether a pupil had completed work.

Miles provided examples of how Chat GPT can write code in seconds that students were given hours to complete. He asked, ‘is this the end for coding?’ and showed how ChatGPT was able to produce a perfect process and code for an OCR assignment that students were given 20 hours to complete in approximately 30 seconds.

Quoting Fahood Manjoo’s opinion piece for the New York Times, he said

“A.I. could transform computer programming from a rarefied, highly compensated occupation into a widely accessible skill that people can easily pick up and use as part of their jobs across a wide variety of fields. This won’t necessarily be terrible for computer programmers — the world will still need people with advanced coding skills — but it will be great for the rest of us.”

AI-generated code is on the way to being firmly part of the programming education landscape, but teachers have not yet adapted practices to overcome the challenges and leverage the benefits of AI. However, teaching subject knowledge will still be important, particularly to write ‘prompts’ and to ‘sense check’ AI generated content.

He agreed with current DfE guidance, that; “It is more important than ever that our education system ensures pupils acquire knowledge, expertise and intellectual capability.”

As well as strong subject knowledge, non-cognitive skills will become increasingly important, as exemplified in the BCS Barefoot Computing model of computational thinking, which emphasises tinkering, creating, debugging, persevering and collaborating.