Schools found to be unprepared for impact of ChatGPT on education

Research conducted by BCS, the chartered institute for IT, found that 62 per cent of professionals believe that chatbots like ChatGPT will make it harder to mark students’ work fairly.

ChatGPT is a large-language model (LLM) which can answer questions in a seemingly natural way and is trained on a massive data set. It has been shown to be able to create passing-grade answers at university level – including passing law exams at one university – but it is fallible. A recent public demo by Google’s own AI service, Bard, produced a wrong answer.

The majority (56 per cent) of the 124 computing teachers in the BCS study did not think their school had a plan to manage incoming use of ChatGPT by pupils, while 33 per cent said early discussions had taken place and a further 11 per cent said a plan was being formed.

Over three-quarters of computing teachers (78 per cent) rated the general awareness of the capabilities of ChatGPT among colleagues at their school or college as ‘low’ or ‘very low’.

Nevertheless, a recent study found that students using the software to complete essay assignments could be risking plagiarism due to the way the AI processes text.

45 per cent of the computing teachers were confident that ChatGPT is a tool that will improve teaching in their school, long-term, by helping to plan assignments and support students with research techniques.

Julia Adamson, managing director for education and public benefit at BCS, said: “Assuming these generative AI programmes remain freely accessible, teenagers are going to use them to answer homework assignments, just like adults will come to rely on them at work.

“Computing teachers want their colleagues to embrace AI as a great way of improving learning in the classroom. However, they think schools will struggle to help students evaluate the answers they get from chatbots without the right technical tools and guidance.

“Calculators used to be banned from exams but are now mandatory. We need to bring machine learning into mainstream teaching practice, otherwise children will be using AI for homework unsupervised without understanding what it’s telling them.

“Another danger is that the digital divide is only going to get wider if better-off parents can pay for premium services from chatbots – and get better answers.”

Concerns have previously been raised about how chatbots could be spreading misinformation through not being able to understand the authenticity of their sources and the broader impact of the technology’s impact on writing-based professions.