Values and Attention in the Age of AI | Charterhouse School

[1] See Trust in Schools: A Core Resource for Improvement (2002).

[2] Richard P. Feynman, “Cargo Cult Science”, Some remarks on science, pseudoscience, and learning how to not fool yourself. Caltech’s 1974 commencement address.

Values and Attention in the Age of AI
Jacob A. Forward
AI Research Associate, Charterhouse School

Generative Artificial Intelligence (GenAI)[1] already gives an impression of omniscience; it now also has a good claim to omnipresence: in the news cycle, in our conversations, in the workplace and, of course, in education. Stories of AI agents escaping their test environments and wreaking havoc online, fears of recursive self-improvement leading to artificial superintelligence and the extinction of humanity, all compete for attention with triumphant techno-utopian tales of limitless power and prosperity. Yet few in this cacophony would grant GenAI omnipotence.[2] For now, it seems, the power remains in our hands.

No sector feels the weight of this responsibility more acutely than education: schools, colleges, academies and universities, all charged with preparing the next generation for a world that is accelerating beyond the pace of understanding. In one of my roles as Director of AIME Education, an AI literacy company, I work closely with schools and colleges, and two challenges keep recurring: assessment, and future-proofing the curriculum.

On assessment, staff are struggling to adapt their teaching around the risk of students using GenAI in their work; there are real concerns about whether any learning is happening; and some teachers, feeling usurped by this new knowledge technology, are questioning their role. Students, meanwhile, rail against the hypocrisy of staff who use AI to mark and give feedback on their work, feel acute fear of missing out, and worry that the top grades will go to those who find ways to use AI. Mutual suspicion is growing, relational trust is deteriorating, and disillusionment with the whole enterprise is doing real damage to the social fabric of school communities. This cat-and-mouse game of deception and detection, played in pursuit of better grades, signals that we are losing sight of the true purpose of education.

The second challenge, future-proofing the curriculum, asks how to prepare young people for the world they will inherit: how to equip them with the skills and ideas they need to live fulfilling lives. Both challenges, I will argue, are at root questions of value.

Shaken awake

The speed and scale of GenAI’s impact over the past three years have sent shockwaves through the education system, and this may turn out to be a good thing. The Czech philosopher Jan Patočka, who suffered political persecution under Czechoslovakia’s Soviet-backed regime, spoke of crises shaking people awake from the sleepwalk of modern civilisation. Such an awakening, he argued, can forge what he called the ‘solidarity of the shaken’, and with it a community of the morally responsible, capable of confronting the truth.[3] Schools are, or should be, communities of exactly this kind. In Magnifica Humanitas, the widely read encyclical of May 2026, Pope Leo XIV reaffirmed the importance of the school in the age of AI as “the place where new generations can learn to seek and love the truth, to reflect on the meaning of life and to recognize the dignity of every person.”[4] I see AI’s impact in similar terms: as a ‘shaking awake’ that can gather a school community around what is fundamentally valuable.

AI as a test of what schools value

What, then, do schools currently say they value? Values operate at many levels in education today: individual, institutional, national (‘British values’) and international (from bodies such as UNICEF or the Catholic Church). The question is whether they are ornamental or inhabited: whether they orient a school’s daily life or simply decorate its website. To find out, I compiled a dataset of over 10,000 English schools and the values they publish on their websites. Identifying and extracting values from these web pages was quite a programming challenge, given their heterogeneity, which limits the accuracy of the results. The sample is also skewed towards schools that have web pages, which not all do. Using the rich data freely available from the government’s Get Information About Schools (GIAS) service,[5] I tested whether school values vary with a school’s setting (urban or rural), its sector (independent or state), the age range and gender of its pupils, its level of deprivation (measured by the proportion of pupils eligible for free school meals) and its region. Despite the methodological limitations a clear high-level picture emerged, and it surprised me. None of these factors predicts much of the variance in school values at all.

Instead, schools appear overwhelmingly to draw their values from the top down, not the ground up. The most common triad, “ready, respectful, safe,” comes from the behaviour expert Paul Dix.[6] Schools also borrow biblical language endorsed by the Church of England, UNICEF’s Rights Respecting Schools, Neil Hawkes’s Values-based Education, and the Department for Education’s ‘fundamental British values.’[7] This suggests that schools espouse values primarily to signal affiliation with missions broader than their own. There is nothing wrong with this, but the disconnect between stated values and the places and people each school serves raises the question of how deeply those values are embodied in day-to-day practice.

Yet if Whitehead is right that “as we think, we live,”[8] then values of some kind are at work in every school’s daily life, whether or not they are the values broadcast on its website. Indeed, values shape our perception so powerfully that the philosopher Max Scheler spoke of Wertnehmung, or ‘value-ception’: we perceive the value of a thing as directly as its colour.[9]

Values and attention

Attention and values are inextricably linked. By ‘value’ I mean not only what we cherish but what we believe to be important. We pay attention to the things that we value, and what we pay attention to reveals what we find valuable. Look at what a school actually attends to each day, and you will see what it really values, which may not be the same as its stated values. To a large extent, our very reality consists of what we repeatedly pay attention to. A person’s values, beliefs, attitudes, and character are embodied in patterns of attention. If I claim to value beauty but never notice beauty, then beauty does not meaningfully exist in my lived world. If I claim to value justice but never attend to justice, then justice is not an operative value in practice. In a sense, then, attention is the enacted form of a value.

Educators may find it dizzying to read, for example, the OECD’s list of overlapping terms for ‘attitudes and values’: beliefs, dispositions, mindsets, virtues, social and emotional skills, ethics, attributes.[10] Dividing the human being into these facets may be useful for measurement and curriculum design, but phenomenologically they all manifest through a more primary process: the selective allocation of attention.

A courageous person notices different aspects of the same reality from a fearful person. A compassionate person attends to suffering that others overlook. An artist sees patterns, colours, and relationships that are invisible to others. The external world may be identical, but the experienced world differs because their attention differs.

Attention, I would argue, is the precondition for all the rest. Modern education may be asking the wrong first questions, “What virtues should students cultivate? What attitudes should they adopt?” when the more fundamental question is, “What are students learning to pay attention to?”

A student who attends primarily to grades, performance metrics and examination requirements inhabits a very different reality from one who attends to truth, beauty, complexity, the suffering of others and the interconnectedness of systems. Over time, they become different kinds of people. This is more than semantics, and neuroscience can offer us some insight.

Two modes of attention

Here the work of Iain McGilchrist is particularly illuminating. A former Fellow of All Souls College, Oxford, consultant psychiatrist at the Bethlem Royal and Maudsley Hospital, and research fellow in neuroimaging at Johns Hopkins, McGilchrist has devoted much of his career to the divided brain, and his account rests on a vast body of neuroscientific and clinical evidence. That evidence shows that each hemisphere of the brain specialises in a distinct mode of attending to the world, and that both are evolutionarily advantageous. The left hemisphere specialises in narrow, focused attention, like a bird picking out grain from grit as it pecks the ground. The right hemisphere specialises in broad, open, vigilant and social attention: the same bird keeping watch for predators and for where it is in relation to everything else around it.[11]

Throughout his work, McGilchrist argues that Western institutions increasingly privilege the left hemisphere’s mode of attention: abstraction, categorisation, manipulation, measurement and utility. He worries that many domains, education included, train people to operate in that mode at the expense of richer forms of awareness.[12] Both modes of attention are necessary, but they disclose very different worlds.

Two ways of responding to GenAI

What does this mean for how schools respond to GenAI? If values and attention are inseparable, and primary in all that we do, McGilchrist’s two modes of attention suggest two possible paths.

Attending to the challenge in the left hemisphere’s mode, we define terms, delineate goals, establish frameworks and committees, write progress reports, update our criteria, and build transparent procedures and governance structures. Even the well-worn formula that GenAI offers ‘promise and pitfalls,’ into which everything can be sorted, is a reductive binary. These managerial superstructures do not, in themselves, change or solve anything; people directing their attention to the problem, alone or in concert, do. The attention, not the framing, is the real work, and we should not confuse the two.

There is a deeper risk in this default approach of managing change through an abundance of procedures and structures, whose very complexity and officiousness lull us into a false sense of safety. Hannah Arendt wrote compellingly about the danger of individuals yielding their agency to institutional directives, surrendering to the reassuring march of abstract processes their moral responsibility to think for themselves.[13]

Attending to the challenge in the right hemisphere’s mode, we encounter a very different picture. Think of a staff room at break: a dynamic, living space, a complex fusion of collegiality, serendipity, discovery and deeper conversation. Its value cannot easily be codified, measured or mandated, but it is undeniable. Like bees returning to the hive, beaks (Charterhouse vernacular for teachers) turn their attention to one another and to the bigger picture. There is space for reflection on individual pupils or on the school as a whole, for sharing difficulties and highlights, and for reorienting the teaching community. This is the right hemisphere’s view: open, social, undefined, sensitive, paradoxical. Yet without some structure, common rules and planning, it too would achieve little. Neither mode is sufficient on its own.

A healthy balance

What is needed, then, is balance. Schools still need standardised testing and ways of measuring improvement; these belong to the statutory environment and to the enduring social contract of education, with its promise of progression to university and careers. It will be interesting to see how far this traditional vision survives GenAI’s impact on the workforce, and whether our thinking about the value of education shifts more profoundly in the coming years.

AI may well liberate us from many bureaucratic burdens, precisely because it thrives where everything is explicit, written down and manipulable. A large language model is, in a sense, the left hemisphere’s mode of attention made machine. For the same reason, adapting ourselves and our institutions too far to suit AI risks excluding the embodied, experiential and tacit sides of life. The deeper danger, McGilchrist warns, is not simply that the left hemisphere’s mode becomes predominant, but that it is utterly convinced of its own completeness. The right hemisphere, open to every perspective, including the left’s, never presumes to see the whole; the left hemisphere believes it alone has everything. A technology that embodies the left hemisphere’s way of seeing can only amplify that dangerous conviction in us, unless we consciously attend to what it leaves out.

In my work as lead tutor on the Generative AI in Business course for Cambridge University Press, I constantly hear how businesses are struggling to automate operations with AI. They are discovering entire worlds of the implicit and the intuitive, what might be described as tacit knowledge, even in the modern workplace. This is hardly surprising. As the philosopher Michael Polanyi put it, “we can know more than we can tell.”[14] This tacit knowledge is often what makes an organisation work at all: the judgement of an experienced colleague, the unwritten understanding of a team member’s strengths, the trust that lets a problem be solved without an explicit procedure. Not everything can be codified, automated, proceduralised or rationalised, because the infinite complexity of real, embodied human social life defies capture in this way. If this is true of the workplace, it is truer still of the school, where so much of what matters, from a pupil’s growing confidence to a beak’s sense of when to push and when to wait, is tacit. Schools are on the side of human beings. They are in the business of helping pupils develop as people.

What changes when attention shifts

What, concretely, would it mean for a school to shift its attention? My thesis is that if schools move beyond the left hemisphere’s view of education to encompass more of the right’s, they will see their core work in a very different light: from the product to the process, from the measurable to the meaningful, from isolated parts to the whole person, and from transactions to relationships. Consider what this means for assessment, for beaks and for students.

Take assessment first. MIT’s recent internal review of AI in teaching and learning offers a telling thought experiment: what if there were no grades? Much of the incentive to use AI to shortcut learning would disappear. As the committee also observes, “if we want students to care about and know how to create and recognize worthy work – work of scope, rigor, creativity and thoughtfulness – quick, high-stakes evaluations embody the opposite of the signal we want to convey to them right now.”[15] Abolishing grades is likely not feasible today, but I think it should become the norm for beaks to explain the pedagogical purpose of every assessment they set. Pupils would then always encounter the reasons for their work, and their attention would be repeatedly drawn to the value of the process, not merely the outcome.

For beaks, the shift is from content delivery to relationship. Those who feel their role is threatened are right to ask the question. Delivering knowledge was only ever part of the job: the rest is modelling learning, care and curiosity as human beings. It is a shift that Charterhouse’s tutor and House system aim at, and at their best achieve.

For pupils, the shift is from performance to growth. A pupil who sees an essay as a product to be submitted may reasonably ask why AI should not write it, whereas a pupil who sees it as a way of learning to think will see that outsourcing it defeats the point. Attention shifts from ‘What will get me the grade?’ to ‘What am I becoming by doing this?’ Across the whole school community, shifts like these may begin to help in repairing the damage AI has done to the relational bonds of trust and shared purpose.

Looking forward

Since joining Charterhouse in September 2026, I have continued the school’s work of engaging deeply with pupils and staff about their experiences and views of AI, and of pioneering what a curriculum could look like in the age of AI. Our focus is on how a school community can learn with AI wisely and together, while taking its positive potential seriously. This positive potential should not be understated. Used well, AI can give every pupil a tireless Socratic tutor, open up knowledge once reserved for specialists, and let curiosity follow its own meandering path across subjects. Personally, I am most excited about AI’s potential to make polymaths of us all.

I have been struck, in my first few weeks at Charterhouse, by a sense of what it must have felt like to work at NASA in the 1960s: the boldness, the capability, the collegiality. I look forward to bringing my experience to bear, and to sharing the findings of our experimentation with the wider education community. In particular, I would like to pose a metacognitive challenge to everyone, from pupils to school leadership: to become more aware of their attention, whether it is captured or directed, what they attend to each day, and the quality of attention they bring to it, with McGilchrist’s two modes in mind.

I close with an irony of poetic proportions. In 2017, a team of researchers at Google published the paper that began it all. It introduced the ‘Transformer’ architecture that underpins today’s large language models and ignited the GenAI boom we are still experiencing. They had made a true breakthrough, discovering how an all-important human faculty could be mimicked by AI, and they duly named the paper after it: “Attention Is All You Need.”[16] For schools facing AI, perhaps it is.