As researchers urge ministers to get artificial intelligence into the curriculum, an Obama-era tech adviser warned the real risk may not be the technology in the classroom — but decision-makers who refuse to understand it.
Typically, the debate over artificial intelligence (AI) boils down to two approaches: either AI is the great liberator or the herald of the robot apocalypse.
This is a framing Professor Beth Simone Noveck knows well, and one she is keen to dismantle.
During a recent visit to Edinburgh to speak at the “Civic University – Democracy, AI, and the Public Imagination” event, she described two typical headlines she sees relating to AI: “They either talk about the robot apocalypse… or there’s the equal and opposite adulatory.
“What’s often missing is what’s in between.”
Prof Noveck — a former US Deputy Chief Technology Officer under President Barack Obama, now a professor at Northeastern University — has spent years in that in-between, which she calls the “messy middle”.
And it is the middle, she argued, where decision-makers and educators need to be working.
Professor Beth Simone Noveck, an expert in AI and a former tech advisor to President Obama, sees AI as a tool to be used alongside humans and not a replacement. (Image: University of Edinburgh)
Shortly after her visit, the University of Edinburgh released a report calling on the Scottish Government to “provide sustained investment in data and AI skills as an industrial strategy commitment rather than a discretionary education spend — building the human capability to apply AI well, not just the infrastructure to run it.” In short, leading AI researchers are calling on ministers to find ways to embed AI into the school curriculum.
The case is partly pastoral: Training teachers to use and teach AI will equip young people not only to use it in the workplace but also to understand its risks and impacts.
The case is also economic. The report cites research showing that 62% of Scottish business leaders rate their organisations’ data and AI literacy as moderate or low — a skills gap, its authors argue, that leaves the country unable to use the infrastructure it builds. Businesses increasingly expect even entry-level recruits to arrive with a basic working grasp of these tools. But the case is also, in Prof Noveck’s telling, about competence of a deeper kind: AI works best, she insisted, with a human steering it, and you cannot steer what you do not understand.
That is easier to argue in the abstract than to sell to the people most exposed to the downside. There is a powerful fear that AI is here to take our jobs — especially the jobs typically done by young people fresh out of school, college or university, hoping for the entry-level positions that start a career. Prof Noveck recognised that these fears exist and that, in many ways, they are justified.
“AI will be used as the excuse to lay people off,” she said. “To increase the CEO’s pay and increase shareholder value.”
Even so, there are complicating details that bear consideration, even if they do not fully relieve the anxiety of a generation. She referenced a project she helped build in California, shorthanded as AIEP, which uses AI to translate and explain the dense, lengthy documents that set out a child’s Individual Education Plan (IEP) for parents and carers who are not proficient in English.
She described how it was built with parent ambassadors and engineered so that a document about an individual child downloads only onto the parent’s own computer, “not in some cloud somewhere.”
A report from the University of Edinburgh has called for AI to be embedded in the curriculum to meet the needs of businesses and prepare young people for the modern workplace. (Image: Gordon Terris)
When pressed, she admitted that, of course, there are human translators who could do the job, but the truth is they are not always where they need to be at a given time to help a desperate parent. That capacity may eventually come, and she argued it should, but the investment in human translators will be costly and slow.
In the meantime, the personal cost would fall on potentially hundreds of families and young people who might face disadvantages they did not truly have to.
Read more
Even in examples like the AIEP tool, the point Prof Noveck was most keen to stress is that AI is not, in itself, a perfect solution. She sees it working best alongside a human.
She recalled a description someone once shared with her: that AI is “like 1,000 interns on cocaine.” The interns will be energised and motivated, and they will produce lots of content — but giving them free rein is bad for business.
In Scotland, a recent innovation out of the University of Glasgow turned heads when developers announced they were trialling an AI tool that records and transcribes classroom discussions, converting them into reviewable packages teachers can work through — with the AI’s help — to pinpoint critical moments in the conversation and consider how they might reshape a lesson.
The tool, called a Deliberative Instructional Agent, or DIA, is the product of four years’ work by Dr Thomas Cowhitt, a teacher-educator at the University of Glasgow, and Dr Matt Gibson, Rector of The Glasgow Academy. It records, transcribes and attributes who says what during a lesson, then — within roughly 100 seconds of the bell — hands the teacher a dashboard flagging four to six “critical moments”, the pivotal junctures where a teacher’s response shaped the direction of the class
Returning to Noveck’s two typical views of AI, there will have been many who read about this development and saw only either “AI will record pupils in class” or “AI tool gives teachers instant feedback.” In reality, the tool sits squarely in the messy middle. It is not preloaded with pedagogical best practice, nor designed to steer a teacher towards any particular method — and, the developers stressed, it is not a surveillance tool.
“EdTech tends to promise all these transformations, and generally falls short,” Dr Cowhitt said. “A lot of educational technology is developed by software developers without the input of educators, and that is the opposite of what should be happening.”
Dr Gibson put the distinction more sharply: “The power is not in this AI doing clever things for you. It is about channelling human reflection.”
It is, in some ways, Prof Noveck’s principle in practice, with the assistant in the background and the human in charge.
The truth is that AI is already not just in our lives but in our schools — and in many ways has been for some time. Strip away the marketing, Prof Noveck argued, and we will see that generative AI is a relative of tools we long ago stopped fearing: the spellchecker that finishes your sentence, the email that suggests its own reply, the dictation tool on a phone. All of this, she points out, is the machinery learning patterns in language, long before anyone called it AI.
The choice to exclude it may, in any case, already have been taken out of our hands. When I asked what she would say to the young professionals, academics and school pupils who see AI as the death knell for careers that looked promising only a few years ago, her answer came as close to a warning as she offered all afternoon: start learning about it, and do your best to make sure it is used for the right reasons.
“There’s a real feeling that we got this wrong with the internet, we got it wrong with social media,” she said. The instinct then was to ban, to look away, to treat the thing as a problem to be kept out rather than a tool to be understood — and as a result they ran rampant.
She wants to make sure that same mistake isn’t repeated.
“That which we don’t know is always what we’re scared of, and then we make bad decisions about how to use them.
“Our kids use these tools now. For us as adults to put our head in the sand and just say ‘no, this is bad’, means we make bad decisions about how to use them in schools.”
This website and associated newspapers adhere to the Independent Press Standards Organisation’s Editors’ Code of Practice. If you have a complaint about the editorial content which relates to inaccuracy or intrusion, then please contact the editor here. If you are dissatisfied with the response provided you can contact IPSO here
© 2001-2026. The Herald is owned and operated by Newsquest Media Group Ltd, an audited local newspaper network.
Visit newsquest.co.uk to view our policies, terms and legal agreements.
The Echo Building, 18 Albert Road, Bournemouth, England BH1 1BZ. Registered in England & Wales | 01676637
Data returned from the Piano ‘meterActive/meterExpired’ callback event.
As a subscriber, you are shown 80% less display advertising when reading our articles.
Those ads you do see are predominantly from local businesses promoting local services.
These adverts enable local businesses to get in front of their target audience – the local community.
It is important that we continue to promote these adverts as our local businesses need as much support as possible during these challenging times.

Leave a Reply