THE impact of poverty and disadvantage on learning and attainment in Britain today is well documented. The scale of child deprivation equally so, the Joseph Rowntree Foundation predicting the number of children in poverty to rise to 4.8 million by the close of this parliament.  

The last years of Tory government were increasingly punctuated by reports of worsening educational inequalities and widening attainment gaps.

The Education Policy Institute (EPI’s) last annual report showed gaps at record levels for reception children with SEND (Special Educational Needs and Disabilities) and an attainment gap of over 19 months for pupils sitting their GCSEs.

When it comes to the emergence of AI into the EdTech market of British education it is essential to remind ourselves of the depth of the crisis of inequality writ large in the lives of our children and the schools they attend.

It’s this context that, without intervention, will dictate everything – from its uses in the classroom to its impact on attainment gaps. AIEd (Artificial Intelligence in Education) could well become the tool to hammer home the disadvantage gap and deepen the gulf.

It’s a dangerous context, where every marker of the crisis our state schools face – the high class sizes, absence rates, unmet SEND need, an unwieldy and non-inclusive curriculum warped by high stakes testing and accountability structures and the loss of professional autonomy, the haemorrhaging of experienced staff and shortage of subject specialists — all of this — driving increasingly unbearable levels of workload. These are the drivers already determining how AIEd will work in schools.

The market is already primed and geared towards using AI as a cheap fix for each.

Take some of the individual digital learning platforms with their repetitious formatting, constant assessments and acknowledged risk that they are so isolating and dull to learn from they risk even more pupils voting with their feet, but for which demand is expected to grow as they can replace the missing subject specialists.  

Or some of the AI-powered marking tools which claim to be able to assess a whole class’s writing output instantaneously, but which deprive each child’s work of any relevant audience and purpose, of any actual meaning, and reduces it to a spelling and basic grammar grade.  

Or an AI-powered coaching tool reliant on a thin diet of two-way videos bordering on surveillance, which would replace the need for specialised and trained mentors for those schools and trusts lacking staff and with tight budgets, and become the main guide to the vulnerable trainee and early career teacher and shape their emerging pedagogy and understanding of their practice.

Or even some of the more apparently benign AI-ready lesson plans and planning tools. Based on bought-in or in-trust standardised curricula or to blend with micro-managed teaching scripts and stripped of professional input, the insight of the pedagogical and subject specialist, the empathy of the responsive educator to their own community of pupils and driven by time and workload pressures – even these risk worsening the standards and quality of education many children would receive.  

We know there is a clear correlation between schools in disadvantaged areas and pupil absence, SEND need, staff turnover and subject specialist shortages.  

The real danger is these are pressures to use AIEd in all the wrong ways and for all the wrong reasons, and in so doing bake in the disadvantage gap by creating the conditions for more disillusioned and unhappy children, more isolated and stressed staff with eroded autonomy, eroded knowledge and skills, working outside of their subject specialisms and expected to cater for some of the highest class sizes and SEND needs in recent record with inadequate resource because “there’s an AI for that.”
  
Then there is the mirroring of significant inequality in the current use of AI with The Alan Turing Institute’s study most recently showing 52 per cent of pupils attending private schools report using AI, as opposed to only 18 per cent of children in state schools. It also records a similar disadvantage gap in use according to household income.

It’s notable that many independent schools are centring technical knowledge equally with creative uses and critical analyses of AI, envisioning uses which confer choice to the teacher as a professional specialist, able to maximise the small class size to offer genuine learning choices and exploit its many progressive and motivational opportunities. 
And also conferring responsibilities and choice to the child, in terms of an awareness and an ability to analyse both how and what they study.  

They seem to have well absorbed a lesson from the past: that perhaps it is preferable to use the technology, rather than to allow the technology to use you.

They offer models of use which all state schools and colleges should be enabled to note, learn from and even exceed.  

Yet that requires the legacy of years of austerity to be addressed.      

If it is not, if AI is encouraged to operate unchallenged and under-regulated by government and unmediated by the collective voices of education workers, we risk not only the sources of the crisis in schools remaining unaddressed, but also the dangerous escalation and compounding of every symptom of that crisis.

We risk alienating and disadvantaging both educator and child.

It cannot be that it is left to where each setting sits on the lengthening continuum of wealth and disadvantage in Britain that determines the use and impact of AI on our education system, our children and the future fabric of our society.

Nicola Sarah Hawkins is a member of the NEC (national executive committee) from NEU District 1 (north-east of England).