In 1999 Napster was the first of many ‘pirate’ websites to give free access to the majority of all the music ever recorded. Over the subsequent decade the $30 billion global market for recorded music shrank by about a third (in inflation-adjusted terms) before slowly returning to growth through streaming models such as Spotify.

Embedded assumptions about the fairness of IP protection came under attack during that stressful decade. My then employer, EMI Music, sponsored my part-time doctoral research to explore how we might adapt to the disruption caused by the extraordinary phenomenon of peer-to-peer filesharing. The research piqued the interest of Macmillan Publishers and I switched jobs into a sector whose copyrights were similarly threatened by shifting cultural attitudes and technology evolution.

My research led me down some fascinating post-structuralist pathways. It introduced me to Roland Barthes’ Death of the Author (1967) and Michel Foucault’s What is an Author? (1969) making me question my simplistic views about the primacy of an author's intent; about plagiarism and the fairness of copyright law. 

Words are never entirely fresh or neutral. They are endlessly recycled through the context,  interpretive bias and emotional baggage of each writer, reader, and listener. This is rather neatly captured by Dennis Potter’s quote above.

So what precisely is the moral claim of an author (droit moral d’auteur) to demand that their words be attributed to them in perpetuity? 

In the tinderbox of revolutionary France when the concept first emerged, ‘owning’ your words was less about economic protection and more a matter of honour. Owning ‘up’ to what you had written carried serious personal risk. 

In the twenty-first century, when truth is no easier to discern and lies circulate with impunity, we could learn something from eighteenth-century principles of moral authorship. Through attribution, we should at least try to acknowledge how our own words have been shaped by others. It may be our best strategy for the survival of our species.

In the book produced from my research (Patrons, Curators, Inventors and Thieves, Palgrave, 2014), I dug deep into the fiercely contested history of copyright and piracy. I traced how competing metaphors of creative output had been manipulated for political and economic ends:

• the birth metaphor, the brain-child of authors, whose original work must be protected as an economic incentive for creators (and for their patrons and publishers)

versus 

• the derivative metaphor of standing on the shoulders of giants: prior creators who themselves might not have been so creative had they not had free access to what went before. 

Would we worship Shakespeare quite so much had he not plundered Holinshed, Marlowe, Greene, Kyd, and others? In the parallel process of musical composition, would we have Mozart if Bach, Haydn, and Handel had reserved all rights to their work?

One conclusion was that some relaxation in IP protection could benefit the economy. Open Access (OA) and Creative Commons (CC) already pointed the way, calling for less use of the blanket restriction of ‘© - all rights reserved’, a lazy form of corporate protection that hides or restricts access to vast amounts of scientific and creative output. 

By contrast, CC provides licenses to creators to share their work for public benefit whilst optionally reserving ‘some rights’. 

For example, I sometimes use a Creative Commons BY-NC-SA license (see image at the end of this essay). Others can distribute, adapt and build upon my content for non-commercial purposes, as long as attribution is clear and further adaptations are under the same terms.

Few people fully understand from which sources large language models (LLMs) have been trained. Content is ‘scraped’ from the internet indiscriminately so it is implausible that unauthorised material is excluded. Billion-dollar litigation settlements and pre-emptive blanket licensing deals between AI companies and content owners are evidence of copyright transgression. Precise attribution is often impossible and most authors will struggle to receive compensation.

My helpful chatbot (who invariably praises my intelligent questions) has been rather coy on this subject:

Me  - What evidence is there of LLMs being trained using copyright material, e.g. content behind paywalls, either legitimately or illegitimately?

Gemini - The system was unable to answer

Me - Why can you not answer this important question?

Gemini - The system was unable to answer

If non-attributed ‘knowledge’ becomes the norm, is it reasonable to expect young people to develop the citation disciplines of previous generations?  

Yes it is. Caring about authorial history and the relationship between knowledge and power is a quality-assurance guardrail in the discipline of critical thinking. Ceasing to care may accelerate our journey to a dystopian fate of being cognitively-diminished slaves to an all-powerful ‘singularity’  known as artificial general intelligence (AGI). 

This risk must be considered alongside the risk of humans (especially young people) not evolving co-intelligently with AI. The longer we avoid building rigorous co-intelligence processes in our schools, the more likely it is that cognitive surrender (the treacherous allure of learning shortcuts) will become habits embedded through shadow adoption of AI. 

This is a common response these days to diminishing appetites for reading, exacerbated by AI shortcuts. I must be honest - I mastered a pre-AI version of this shortcut in my first year at university (1983). Every literature student  has to make a trade-off between depth and breadth in reading core texts and literary criticism. So let’s not jump to prejudicial conclusions about student AI usage.

Surely what counts is evidence that learning and critical thinking have occurred? This includes the discipline and humility of attribution, authenticating the sources that have helped form the learner’s insights.

The golden age of the 'linear literary mind' refers to three centuries when being able to do alone what LLMs are now doing for us was the route to career success. 

The burning question is: how much should we care about losing our disposition to read, evaluate and write long-form text without help from technology? A similar question applies for composition in music, film and the arts generally.

Today the heavy-lifting of text production (including synopsis for readers) is increasingly performed by generative AI which is often evaluating and summarising the very content it produced in the first place. Why do we accept this irony?  Because LLMs feed our self-centred need to be ‘authors’. This produces oceans of mediocre text which cannot be digested without AI support. As uncurated text competes with traditionally published text, can we trust AI to be a reliable curator and abbreviator?

Neil Postman’s much-cited preface to his 1985 book Amusing Ourselves to Death reminds us of earlier warnings about the decline of reading with the advent of technology, i.e. the television. Only a two-minute read, he brilliantly summarises the astonishing prescience of Aldous Huxley’s Brave New World (1932) in relation to George Orwell’s 1984 (published in 1949):

‘What Orwell feared were those who would ban books. What Huxley feared was that there would be no reason to ban a book because there would be no one who wanted to read one.’

In Brave New World, John ‘the Savage’ meets a tragic end. He is unable to reconcile the Shakespeare-inspired imagination in which he’s raised himself with the lazy, trivial, distracted, narcissistic and narcotic reality of the new World State. A world where (as Postman puts it):

‘people have come to love their oppression, and to adore the technologies that undo their capacities to think.’ 

Almost a century after Huxley published his vision, is there a stronger argument for the ongoing importance of studying literature?

* A term promoted by James Marriott’s book: The New Dark Ages - The End of Reading and the Dawn of a Post-Literate Society, Bodley Head (September 2026). 

We do not expect learners to cite all sources of their learning: e.g. from their textbooks, classroom handouts or insightful phrases their teachers or fellow students might have expressed. If retained, students will pass off such wisdom as their own. Repackaging existing knowledge is, whether we like it or not, the core of our assessment system.

It is good to remember this when worrying about students using AI to help them learn.

Some might see AI as encouraging students to ‘cheat’. Such a response is short-sighted.  AI is a force for neither good nor ill. That judgement depends entirely on whether the student has integrity as a learner, whether they have been trained to engage with AI responsibly, to develop good iterative questions that scaffold their learning, and to acknowledge transparently how they have used AI. 

My more emphatic point is that, used well, generative AI might just be the bespoke Socratic mechanism through which some learners, previously excluded from the academic process, can now engage on their own terms, including orally. If schools are to counter Huxley’s fear of technologies undoing our capacity to read, write, listen, speak and think, we must wrestle and shape AI, not avoid it.

To those pioneering schools using AI in the classroom, I would urge you to  promote the discipline of attribution as a key element of AI literacy:

• encourage student discussion of the pros and cons of recognising intellectual property

• ask students - why should we attribute the sources of what we write?

Try starting with the moral purpose of Creative Commons. You might be surprised how a lively debate will emerge about what human creativity and authenticity really mean. What do fair and reasonable recognition and reward look like for artists and innovators? 

It can open up many inter-disciplinary themes, including conversations about enduring human values, skills and careers in an AI future.

The alarming decline in students choosing to study English at A Level and university is unlikely to be reversed without major reform of the obsolete GCSE, including serious accreditation status for oracy skills. That is another topic, but my related concern is that the humanities in general are wrongly perceived as less relevant in the age of AI. 

The emerging demand is for creativity and rigour in co-intelligent evolution with AI to reap its many rewards, within governable risk parameters. For as long as humans can maintain the upper-hand over AI, asking the right questions and authenticating the answers must be the focus.

Strategic and critical thinking underpin all of this. Thinking that relies on detecting and challenging (or exploiting) the often hidden, seductive and historically-rooted power mechanisms of language. Demonstrating the epistemic* origins of knowledge - from whose mouths, pens or keyboards it has evolved - is an enduringly valuable discipline. If we want humans to remain at the top of the evolutionary tree we must treasure the practice of author attribution.

The study of the humanities can be the best foundation for subsequent pursuits. As an English graduate turned CFO, I feel sure about this. It is the message I try to get across to young people who want to study English, but lack confidence about its rewards.

*The term ‘epistemic atrophy’ has emerged, described by Lucy Gill-Simmen as the slow erosion of our willingness and capacity to undertake the more demanding work of understanding what is real. This includes self-awareness of the mechanisms we often use subliminally to trust or mistrust sources of knowledge.