Schools 'desperate' for clearer AI guidance

Teachers are “desperate” for clearer guidance on how AI should be used in schools, MPs have been told.

Pip Sanderson, director of teacher development at the National Institute of Teaching, told the Commons Education Select Committee that school leaders were struggling to respond to the rapid growth of AI.

She said feedback from leaders attending AI training was often: “This is all great and really interesting, but what do I do?”

Darren Northcott, national officer at the NASUWT teaching union, said Department for Education guidance had been “quite laissez-faire”, and schools want “much clearer rules of the road” on how AI should be used for teaching and learning.

Ms Sanderson called for AI training to become a statutory entitlement for trainee teachers, early career teachers and those taking national professional qualifications.

“They are desperate for really clear guidance and advice around what to do,” she said.

Questions over AI impact on workload

During the session, the committee also heard doubts about the government’s hopes that AI could reduce teachers’ workload and help tackle recruitment and retention problems.

“The jury is very much still out” on whether AI would deliver sustainable workload reductions, Mr Northcott said.

He added: “I don’t think the case is proven that the increasing use of AI is going to reduce workload in a way that supports addressing some of those retention challenges that we face.”

Professor Rebecca Eynon, a researcher at the University of Oxford, said AI could instead “reconfigure” teachers’ work, with time saved on some tasks offset by checking its outputs or managing the technology.

She added: “I’m not sure it’s ever going to save anybody any time.”

Warning over disadvantaged pupils

Experts also warned against relying on AI tutors to support pupils who are struggling academically.

Professor Eynon said research in secondary schools had found personalised and adaptive systems could be “very demotivating and kind of soulless experiences”, particularly for students with special educational needs and disabilities or those struggling academically.

Schools serving more disadvantaged areas were relying more heavily on technology because they had fewer teaching assistants or faced teacher shortages, she said, potentially “compounding the problem”.

Ms Sanderson said there was no “magic bullet” that could automatically provide good teaching for disadvantaged pupils, and she warned against replacing teacher relationships with “a screen AI tutor relationship”.

She also said there was a danger of prioritising the wrong things in addressing educational disadvantage: the focus should be on getting more high-quality teachers into the classroom, “rather than trying to use AI chatbots to solve that problem”.

Concerns over coursework

Sir Ian Bauckham, chief regulator at Ofqual, said some extended writing assessments completed with little supervision were now “exposed to risk” from students using AI.

This could compromise the extent to which assessments “really tell us the truth” about what students know, he said.

Sir Ian said extended writing coursework should be avoided “wherever we can” and adapted where it could not be removed.

He also said that “we need to send very strong messaging to students about the consequences and sanctions of malpractice involving AI, which is effectively cheating”.

Colin Hughes, chief executive of the AQA exam board, said teachers were facing an additional burden in determining whether students had used AI.

AQA was exploring the use of AI to “mark the markers” by monitoring human examiners rather than replacing them, he added.

Pisa findings on AI chatbots

Recent analysis by the Programme for International Student Assessment (Pisa) also found that “AI chatbots are now routinely used for schoolwork in many countries”.

Pisa, whose data draws on 91 countries and economies in the Organisation for Economic Co-operation and Development (OECD), found that “students who report using AI chatbots get lower scores in science, on average, than those who don’t use AI”.

It added that “students who don’t use AI outperform those who do, on average, by around 20 points...equivalent to being around one year of schooling ahead of their peers”.

However, the Pisa analysis stressed that “these findings do not necessarily represent a causal relationship”, adding: “Lower performance outcomes could be linked to AI use or to other reasons, such as struggling students potentially being more likely to rely on AI.”

The purpose and frequency of AI use also appeared important, with the Pisa analysis saying: “For example, students who report using AI chatbots to help them learn once or twice a week perform at a similar level to students who report not using AI at all, on average across the OECD. This reflects the notion that digital tools can support learning when used purposefully and in moderation.”

In contrast, however, “overuse or a lack of guidance can lead to distraction, limiting meaningful engagement and negatively affecting learning outcomes”.

Pisa also raised concerns about “a potential AI divide”, with advantaged students appearing to have more opportunities to use AI while also learning in school how to assess the quality of AI-generated information.