Why the ￡3 million AI “content store” for schools could just speed up teacher stress

“There must be an ‘industrial revolution’ in education, in which educational science and educational technology combine to modernize the grossly inefficient and clumsy procedures of conventional education.” Sidney Pressey wrote that in 1933 after the commercial failure of his automated testing and grading device that “does away entirely with the drudgery of grading tests and quizzes”. The labour-saving revolution in education never happened. Now, the UK government announces a “new project that will bring teachers and tech companies together to use trustworthy AI tools that can help mark homework and save teachers time”. The project will, it claims, harness the power of tech to ease the pressures and workload burdens on teachers – pressures that have increased considerably over the ninety years since Pressey’s New Automatic Testing Machine.

The ￡4 million government project will spend ￡3 million on an online store of documents – including curriculum guidance, lesson plans and anonymised pupil assessments – for tech companies to train their AI tools for teachers. An additional ￡1 million will be awarded for ideas to put the data into practice to reduce teacher workload. The refrain of increasing teachers’ efficiency through AI is echoed in recent report from the DfE that claims educators are already saving time on lesson planning and report writing, improving their work-life balance.

Spending ￡4 million of Government money on helping tech companies develop AI tools for automated marking and lesson planning will not reduce pressures on hard-working teachers, it will most likely just speed up their stress. Let’s think this one through.

Wisdom not data

Teachers spend over half their time not teaching students - ?in planning lessons, creating content, grading assignments and doing admin work – so it make sense to reduce that burden. AI tools can already help teachers find content, plan lessons and explore new ways to teach. ??The tools could be improved by basing them on best practice in lesson planning and evidence-based methods of teaching and learning. I have to declare an interest here – I’ve developed a ChatGPT pedagogy expert for teachers based on my book Practical Pedagogy: 40 New Ways to Teach and Learn.

Offering teachers powerful tools to extend their skills and ease their admin is a good thing. But it doesn’t require a ￡3 million database. AI chatbots have already been trained on billions of pieces of data. What’s needed is to give the bots specialist training in the science of learning and effective methods of teaching for differing students and contexts – wisdom not data. A partnership of educators with AI companies could produce better-informed and more appropriate tools for teachers. But that’s not what is proposed.

As Schools Week notes, the ￡3 million “content store” will be for tech firms to train generative AI tools that will then be sold to schools. Government ministers hope it will result in more accurate software that boosts teacher efficiency. The flaws in this mission need to be unpicked.

Technological determinism

First, there’s an underlying technological determinism – the idea that an injection of technology into an educational setting will improve teaching and learning. That has been debunked in many well-argued studies. As the Education Endowment Foundation states: “to improve learning, technology must be used in a way that is informed by effective pedagogy.” The urgent need for AI education companies is not more data but better understanding of pedagogy and how to align their tools with a variety of practices.

An AI arms race in education

Second, an emphasis on boosting efficiency could have the opposite effect. Giving teachers software for marking homework could just lead to an AI arms race. Students use increasingly sophisticated AI to write assignments. Teachers use AI to assess them. Nobody learns; nobody gains – apart from the edtech companies that supply these tools. Teachers know their students, in part, through marking and commenting on written assignments. Handing marking over to an AI loses that connection.

Variety and Professionalism

Studies with previous generations of personalised tutoring for students and automated admin systems for teachers have shown that these technologies typically increase, not decrease, workload. They require teachers to incorporate new technology into their already-full lives and can undermine their professional autonomy and expertise. Leon Furze points out that teachers don’t want AI telling them how to plan lessons faster. ?Rather than a centralised AI education database and a focus on efficiency, the project could enable variety and professionalism – helping teachers choose the best tools for their working lives, to benefit their own students and enhance their expertise. If educators, educational technologists and AI companies can work together to design new pedagogy-informed tools and resources that build on classroom practice and teacher professionalism that will be money well spent.

[The project involves a partnership with The Open University, which is sharing relevant learning resources. I have no involvement with that partnership.]