EP26: Principles before Platforms: Safe Human Centred AI in Education with Sapio

From Teaching to Digital Education

Laura Knight spent 20 years working as a teacher across both the maintained and independent sectors before moving into consultancy. She now runs Sapio, which focuses on digital strategy and technology in education.

Her work spans schools, trusts, universities, publishers, edtech companies and governments, covering everything from devices and procurement to policy, digital wellbeing and curriculum design.

AI has increasingly become central to that work, but Laura's approach is broader than simply adopting the latest tools. Her focus is on helping education use technology well while keeping people at the centre of innovation.

She describes herself as a "good kind of troublemaker" — someone willing to challenge existing thinking and push conversations towards more useful approaches.

AI Adoption Is Growing from the Ground Up

Unlike many businesses, where AI adoption can be driven from the top down, education is often seeing the opposite pattern.

Teachers and middle leaders are experimenting with AI independently, driven by professional curiosity and a desire to discover what the technology can offer their classrooms. This experimentation can sometimes happen before schools have established formal policies, security protocols or guidance.

That creates a difficult dynamic. Innovative teachers can be seen either as pioneers or as "rebels" who need to be controlled.

At the same time, students are experimenting with AI independently of their schools. Many young people may encounter AI through platforms such as Snapchat before they have any structured education about what AI is or how it works.

This means that schools can find themselves trying to establish rules around technology that students have already been using in their everyday lives.

Balancing Innovation with Safeguarding

Laura believes that caution from school leaders should not simply be dismissed as resistance to innovation.

Teachers have a responsibility to protect young people, particularly when it comes to data and privacy. AI platforms can lack transparency around what happens to information entered into them, creating potential risks when sensitive information about students is involved.

The distinction is important. Experimenting with AI to generate a chocolate cake recipe presents very different risks from entering sensitive information about a child.

The challenge, therefore, is not choosing between innovation and caution. It is finding ways to balance the two.

Laura points to recent Department for Education guidance for teachers and school leaders as an important step towards giving schools a clearer framework for safe and responsible AI use. However, she argues that guidance cannot simply attempt to prescribe individual tools because the technology will evolve too quickly.

Principles Rather Than Products

The speed of AI development makes long-term, tool-specific policy particularly difficult.

A school that builds its entire approach around a particular version of a generative AI tool risks becoming dependent on technology that may quickly change or become obsolete.

Instead, Laura argues for starting with core principles and then choosing tools that align with those principles. This means considering questions around privacy, security, interoperability, data governance and educational purpose before becoming attached to a particular platform.

She describes the ideal approach as an ecosystem rather than a single technology stack: different tools should be able to coexist, with interoperability as a basic requirement.

For education, this means "treading lightly" rather than putting all of its eggs in one technological basket.

Helping Young People Understand What AI Really Is

AI literacy is about more than knowing how to write an effective prompt.

For Laura, one of the most important responsibilities of education is helping young people understand what AI actually is — and what it is not. She highlights the concept of "magical thinking": when people do not understand how a technology works, they can begin to perceive it as something almost magical.

This can become particularly problematic for children. A young person encountering AI before they have developed a strong understanding of technology may struggle to distinguish between genuine intelligence, statistical prediction and systems designed to mimic human communication.

Laura argues that education needs to demystify AI and build explainability into the way young people encounter it.

This includes understanding concepts such as consciousness, sentience and what it means to be human. The ability of an AI system to produce a convincing human-like response does not make it human — but increasingly sophisticated systems can make that distinction difficult to recognise.

AI Literacy Should Sit Across the Curriculum

One of the biggest curriculum challenges is deciding where AI education belongs.

Laura argues against treating AI as something that should simply be added to computer science. AI is relevant to English, mathematics, science, citizenship, ethics and almost every other area of learning.

Instead, she advocates for AI and digitcal critical literacy as a horizontal capability that runs across the curriculum. That includes understanding how technologies work, but also questions around integrity, trust, truth, plagiarism and democracy.

Students need to understand not only how to use AI, but how to evaluate what it produces and how its use affects the way they learn and participate in society.

The Paradox of Learning with AI

Perhaps one of the biggest challenges is that effective learning is inherently uncomfortable.

Learning requires students to encounter things they do not understand, struggle with them and gradually develop their knowledge. Yet modern digital life has conditioned many people to use technology to remove discomfort.

Laura describes the smartphone as having become a form of digital self-soothing: when people are bored, frustrated or uncomfortable, they instinctively reach for a device. AI can intensify that tendency. If a student encounters a difficult mathematical problem, the easiest response may be to ask AI for the answer rather than work through the difficulty themselves.

This creates a fundamental tension: how can students use AI as a learning tool without outsourcing the thinking that learning is supposed to develop?

For Laura, the answer requires more than technical skills. Creativity and digital literacy need to be accompanied by self-regulation, self-awareness, integrity and a strong moral compass.

Literacy, Numeracy and Digital Skills

Laura argues that AI education should be built on strong foundational skills rather than replacing them. Literacy, numeracy and digital literacy form an important trio, supported by wider character education.

Mathematics remains particularly important, not simply because students may eventually need technical skills for programming AI, but because foundational numeracy helps develop the confidence and understanding needed to engage with technology critically.

The wider objective is to ensure young people become creators and curators, rather than simply consumers of technology.

Supporting Teachers to Experiment Responsibly

Teachers are at the centre of AI adoption in education, but they need support to experiment effectively.

For Laura, the first priority for technology companies should be listening to educators rather than arriving with predetermined solutions. She is particularly critical of technology companies that attempt to solve educational problems without first understanding what those problems actually are.

A school struggling with attendance, parental engagement or unreliable infrastructure may have little interest in an AI tool designed to automate another task.

The starting point must therefore be purpose: what problem is the school trying to solve, and can technology genuinely help?

Building Trust into EdTech

Technology companies also have a responsibility to make adoption easier for schools.

Laura argues that schools should not have to spend significant amounts of time investigating whether a product meets basic requirements around GDPR, safeguarding, parental policy and age appropriateness. Products should make their approach to trust, accountability and transparency clear from the beginning.

She also sees an opportunity for technology companies to contribute more directly to the education ecosystem — for example, through teacher professional development, workshops, internships and work experience.

Opening up those pathways could help young people understand the technology industry while giving teachers greater confidence in emerging technologies.

Raising the Floor Across the UK

Looking internationally, Laura argues that the UK should be more ambitious about developing AI and digital skills. Rather than relying on charities and enthusiastic individuals to fill gaps in provision, she believes there needs to be greater investment in raising the baseline of digital capability across the population.

This is not simply an educational issue. It has implications for employability, productivity and the UK's future economy.

There is also a growing disconnect between the skills young people develop through traditional education and the skills employers need. Closing that gap requires closer collaboration between education and industry.

Creating New Routes into Technology

Laura sees potential in alternative routes such as apprenticeships and Extended Project Qualifications (EPQs).

She points to the example of a cybersecurity EPQ developed in response to a significant shortage of candidates in the sector. Such programmes can create a pipeline between education and industry while giving students an opportunity to explore areas of interest before committing to a university pathway.

She argues that similar approaches could be applied to AI and digital literacy.

There is also scope to work through existing youth networks — from cadets and scouts to guides and other organisations — rather than assuming that every new skills programme needs to be built from scratch.

The objective is to reach young people wherever they already are.

Looking Beyond the Traditional University Pathway

Laura also questions the assumption that university degrees are the primary route into technology careers.

Young people often have to make subject choices at a relatively early age that determine whether they can later pursue computer science or related degrees. This can exclude people who discover an interest in technology later. Apprenticeships, flexible qualifications and opportunities for career changers can provide alternative routes into rapidly evolving sectors.

This is particularly important for AI, where the skills required are changing too quickly for education systems to rely exclusively on traditional qualification pathways.

Keeping Humans at the Centre

Looking ahead, Laura is most excited about the creative potential of AI.

Rather than focusing solely on automation, she sees opportunities for educators to use AI to explore new ideas, solve problems creatively and collaborate in ways that enhance professional judgement. That human-centred approach runs throughout her thinking about AI in education.

The challenge is not simply to introduce more technology into classrooms. It is to ensure that technology strengthens learning, creativity, community and human agency.

As AI continues to evolve, education will need to evolve with it — but without losing sight of what education is ultimately for.