Who Killed the Author?

A Short History of Publishing Panics - and Why This One Is Different

In 1495, the world cracked!

Gutenberg’s press used paper from pulped rags to mass-produce text. Pulping up to make something new...is nothing new!

Imagine shredding every book in a library, boiling the pages down to a grey slurry, and pouring it into a machine that produces - fluently, instantly, for free - something that reads exactly like all of them. That's not a dystopian fantasy. That's a large language model, and the authors? They're in the pulp.

Printer's workshop from Timelines, created and designed by David Salariya, text by Peter Turvey, artwork by John James

From Rag to Data: The Author Gets Lost in the Pulping Process

Pulping is messy. It takes something whole and breaks it down - into fibres, into fragments, into something that can be reshaped.

In the 15th century, books were printed on sheets made from pulverised cloth - rags torn, beaten, and reconstituted into blank pages for the future. Wood pulp came later. But from the beginning, to share knowledge, you had to first break something down.

Books have always been trouble! Almost as soon as Johannes Gutenberg started producing his bibles in 1454, he called in the receivers. He didn't go bankrupt because of the printing press itself, but rather because of a lawsuit filed by his financial backer, Johann Fust, who demanded repayment of loans and interest before the printing of the Bible was completed, leading to Gutenberg losing his workshop and equipment. 

Today’s Generative AI does something eerily similar - not with rags or cellulose, but to culture. It ingests books, blogs, news stories, songs, and turns all this into word slurry. The results aren’t copies, exactly. They’re composites. Simulacra. Something new - and not entirely benign.

The machine doesn’t understand. It reconstructs.

Somewhere in that process, the creator disappears. The original voice becomes residue.

What we’re facing now isn’t just disruption. It’s dissolution - the creators disapeared.

Pulped by Progress - The Printing Press Made a Mess

Gutenberg’s press had already begun transforming Europe by 1495. Books were no longer the cloistered possessions of monks or kings or chained in libraries. Paper made from rag pulp became the raw material for a revolution.

And what followed wasn’t just a flowering of knowledge. It was exciting, revolutionary and chaotic:

The Reformation, the Thirty Years' War, the shattering of religious and political power -these weren’t caused by belief alone, but by the explosive spread of ideas. The press didn’t just print books. It printed trouble.

Five hundred years later, we’re at another hinge point. The printing press made ideas replicable. Now, generative AI makes them recombinable. And once again, the creators are being left behind.

The New Scribes Are Machines

Generative AI models like ChatGPT, Claude, Gemini and others are not just tools. They are devourers. They ingest everything: books, blogs, news, lyrics, academic papers, forums—even pirated PDFs. From that vast feast of language, they produce new responses—fast, plausible, synthetic.

These systems are trained on human creativity - but increasingly, they respond as if no human ever existed.

The reader types a question. The machine answers. The author is gone.

It's easy to draw a line from Gutenberg’s revolution to this one. But it’s not a clean upward arc. Gutenberg unleashed both freedom and fury. This moment may well do the same.

Pulped Paper. Now Pulped Culture.

Gutenberg’s press didn’t pulp anything directly - but the paper it used was made from pulped rags. Remnants, beaten and reformed into something printable.

Because today, it’s not paper being pulped. It’s culture.

The great LLMs of our age - Large Language Models - reduce books, articles, stories, and songs into fragments. Tokens. Probabilities. What emerges is sometimes dazzling. But it’s built on reduction. A kind of unmaking.

Not quoting. Not preserving. Replacing.

And the creator? Uncredited. Unacknowledged. The voice becomes vapour.

Welcome to Pulp History

I’m working on a new series called Pulp History - which might sound like a punchy title, but there’s more to it than that.

It’s about digging up the buried, the bizarre, the boiled down. All the juicy bits. The strange stuff left behind after the first pass of history.

And it’s not lost on me that “pulp” now also describes what’s happening to creative work in the age of AI: ideas mashed, strained, and spat back out by algorithms.

Call it progress, if you like.

But I see warning signs in the pulp.

When Fair Use Meets Fuzzy Ethics

The legal landscape is shifting fast. The New York Times is suing OpenAI. Universal Music Group is suing Anthropic. Meta is under fire for allegedly training models on pirated books from LibGen.

At the centre of all this? A battle over copyright, creativity, and a phrase with sharp teeth: irreparable harm.

Courts are now being asked to decide: Does training an AI on your life’s work count as theft? If no financial loss can be proven, does it matter?

Some judges say no. If the machine doesn’t quote your book - just learns from it - that’s legal. Even inevitable.

And lurking in the background is a darker possibility: a floated executive order (under President Trump) that would exempt AI companies from copyright entirely, in the name of “national interest”.

The press once broke the power of kings.

Now, AI may break the rights of writers.

The Historical and Contemporary Landscape of Credit Removal

The erasure of creators is nothing new. Throughout history, artists, writers, and innovators have been stripped of recognition for their work, whether through corporate policies, legal loopholes, or sheer indifference to individual contribution. The issue extends across multiple creative industries, from publishing and comics to film, art, and science.

The Real Risk Isn’t Copying. It’s Replacement.

No one’s accusing AI companies of selling photocopies of our books. That would be easier to sue.

The problem is subtler - and far more dangerous.

If a child asks an AI to explain the French Revolution, it may generate a summary indistinguishable from a textbook ,I spent years researching and writing. It won’t quote me. It will replace me.

For those of us who write, illustrate, and design books - especially for children, where clarity and originality matter immensely - this isn’t just theft.

It’s erasure.

What Can Be Done?

It’s tempting to give up. To mutter about unstoppable tides.

But we’ve seen revolutions before. And sometimes, those who adapt early shape what comes next.

We must:

• Protect our names: Visibility is legacy. Make authorship undeniable.

• Prototype, not just protest: Build tools, books, and experiences that use AI on our terms.

• Make ethical noise: Highlight what’s lost when creators vanish from the record.

• Strike fair deals: Partner with platforms - only on transparent, remunerated terms.

This isn’t about fear. It’s about agency.

The Pressure of Progress

Gutenberg changed the world with rag pulp, movable type, pressure and progress.

What he made was messy, glorious - and dangerous.

We’re at the edge of a similar moment.

But this time, the raw material isn’t cloth. It’s us.

The Next Chapter: Standing on Silicon Shoulders

But let’s not be turned to pulp without purpose either.

Every generation of creators has stood on the shoulders of giants. Gutenberg’s press wasn’t welcomed with open arms. It was feared, misunderstood, even condemned. Photography was derided as a mechanical cheat until it became art. InDesign was dismissed by traditional typesetters – yet it gave a whole new generation the tools to make books from their bedrooms.

AI is not the enemy. It’s the next blunt, imperfect, and immensely powerful tool in the creative kit.

Yes, we need regulation. Yes, creators should be compensated when their work is used to train AI models. Transparency isn’t optional. And I fully support campaigns like Make It Fair, which push for micropayments and clear attribution.

But let’s not throw the creative baby out with the synthetic bathwater.

Because while AI may be good at blending data, it still can’t match human taste. It doesn’t know the thrill of discovery. It can’t feel the drama of a story twist or the joy of a well-placed pun. But it can save time. It can help a dyslexic child write their first poem. It can give a solo illustrator access to layout skills. And it can allow a tiny publishing team – or even a single creator – to build something remarkable.

This isn’t the death of creativity. It’s a reshaping.

The creators of tomorrow won’t just be painters, writers, or layout artists. They’ll be curators. Directors. Conductors of powerful digital tools. They’ll need empathy, imagination, storytelling – and a fluency in AI, just as once we had to learn QuarkXPress, Photoshop, or HTML.

Progress is always messy. But it can also be magnificent - if we’re the ones holding the reins.

Quark Was Fine. ChatGPT Is Not: Publishing’s Selective Fear of Technology

When Quark XPress arrived in the 1980s, most designers treated it like sorcery. Gone were the wax rollers, cow gum, scalpels, and Letraset sheets that left your fingers black and your temper short and the air fill of spray glue. Overnight, we could move type around without slicing it first. There were mutterings, of course - claims that design was losing its soul - but within a year even the purists were clicking their way to deadlines. The revolution had been domesticated.

Fast-forward to today, and a new technology is causing panic. Not among printers this time, but among editors. Artificial intelligence - large language models, generative text, whatever label you prefer - has crept into the one corner of publishing that still considered itself sacred: the words.

Suddenly we hear pronouncements about authentic voice and the moral imperative to write “without the use of AI.” It’s a strange line to draw. No one ever insisted we set our own type or mix our own inks. No one objected when page grids were downloaded from the web instead of drawn with Rotring pens. Yet when language becomes programmable, people are alarmed.

What’s really going on is a crisis of authority. Design, production, and marketing long ago accepted automation because those fields could be framed as crafts: technical, iterative, improvable. Writing and editing, however, were ring-fenced as art: ineffable, human, unteachable. When a machine begins to mimic taste, rhythm, and tone, it trespasses on the editor’s identity. The fear isn’t of technology - it’s of redundancy.

Publishing has always survived by absorbing its heresies.The printing press was once accused of ruining handwriting; photo-composition was said to cheapen type; the ebook was going to kill reading altogether. Instead, each disruption widened the audience and forced the industry to evolve. AI will be no different. It will not erase editors; it will expose which ones add real value beyond proofing commas.

The paradox is that the people shouting loudest about “human creativity” are already surrounded by invisible algorithms: spell-checkers, auto-kerners, predictive text, metadata optimisers. The difference is that those tools whisper, while AI speaks back. We can no longer pretend the machine is merely mechanical; it now participates.

Perhaps that’s what unsettles us most: the sense that technology has moved from servant to collaborator. But refusing to engage won’t stop the shift -it will only sideline those who cling to a vanishing purity.

If we survived Quark, we’ll survive ChatGPT. The question isn’t whether publishers should use new tools, but whether they can remain human while doing so: still capable of empathy, judgement, and the spark that no algorithm can fake. The future editor won’t reject the machine - they’ll teach it taste.

A Short History of Technological Panics in Publishing

Or, How We Learned to Stop Worrying and Love the Machine

1470s – The Printing Press

The panic: “It will destroy the art of beautiful handwriting!” cried monks and scribes as Gutenberg’s invention rolled into Europe.Reality: Within a generation, presses were churning out Bibles, broadsheets, and bawdy poems faster than anyone could light the candles to read them.Moral: Every age thinks its version of the pen is the purest.

1814 – The Steam-Powered Press

The panic: “Speed will ruin quality!” Traditional printers feared mechanisation would flood the world with sloppy work.Reality: It did - and readers loved it. Penny papers and affordable books turned literacy into a mass pastime.

Moral: Democratization always looks like degradation to the guild.

1880s – Linotype and Hot Metal

The panic: Typesetters swore that molten lead would replace skilled craftsmen with machines.Reality: It did, but it also birthed the newspaper boom.

Moral: Convenience always wins - until nostalgia arrives fifty years later to declare it beautiful.

1960s – Offset Lithography

The panic: “The end of true printing!” lamented letterpress traditionalists, clutching their forme keys. Reality: Offset made print cheaper, faster, and cleaner - a quiet revolution hiding inside every children’s book of the 1970s.

Moral: Most “new” technologies are just better ways of doing what we already love.

1980s - Desktop Publishing and QuarkXPress

The panic: “Real designers use scalpels!”Reality: Real designers used Quark by Monday. Whole art departments migrated from the lightbox to the Mac Plus.Moral: Once you’ve kerned by mouse, you never go back.

1990s - The Internet

The panic: “No one will read books again!”Reality: Everyone read more — they just stopped paying for it. Moral: Access doesn’t kill appetite; it changes the menu.

2007 – The Kindle & Ebooks

The panic: “Print is dead!”Reality: Print sales dipped, then stabilised; ebooks found their niche. The panic moved on.Moral: Readers care less about format than story.

2020s – Generative AI

The panic: “This will erase human creativity!”Reality (so far): It’s making us redefine what “human” means. Editors are learning prompts the way typographers learned lead.Moral: The tools change; taste endures.

So here’s to the next creative revolution.

Not just pulped by progress… but powered by it.

Pulp History - slice into the juiciest, strangest, and most shocking moments in the evolution of knowledge. This time, I'm asking: When machines become scribes, what happens to the people who once held the pen?

---

David Salariya has been making books long enough to remember Letraset, line art, and when editors had names. His bestselling children’s series are now republished by Bonnier's Hatch Imprint, while machines churn out content in milliseconds. He still believes in human hands, handmade ideas, and the magic that can’t be automated. Visit davidsalariya.com before its fed into a large language model.

When the World First Went to Press

In 1454, a book changed hands without a scribe.Lead met rag pulp, and Europe caught fire. We call it a revolution now. At the time, it felt like a mess.

Gutenberg’s press used paper made from pulverised cloth — rags torn, beaten, and reborn as blank pages for the future. From the first printed Bibles to the Nuremberg Chronicle, knowledge spread faster than anyone could light a candle to read it.

But sharing knowledge meant breaking something down first. Pulping is violent chemistry: you take a whole thing, macerate it into fibres, and re-form it into something printable.

That’s how progress works — it un-makes before it remakes.

Pulped Culture: The Digital Slurry

Today’s Generative AI does something eerily similar — not with rags or cellulose, but with culture itself.It ingests books, blogs, news stories, songs — even PDFs rescued from pirate sites - and turns them into linguistic slurry.

The results aren’t copies. They’re composites.Simulacra. Something new — and not entirely benign.

The machine doesn’t understand; it reconstructs.Somewhere in that process, the creator disappears. The original voice becomes residue.

What we’re facing now isn’t just disruption. It’s dissolution.

Pulped by Progress: How the Press Made a Mess

By 1495, printing had transformed Europe. Books were no longer chained to libraries or hoarded by kings. Rag pulp was the raw material of revolution.

And what followed wasn’t peace. The Reformation, the Thirty Years’ War, the shattering of authority - all fuelled by print’s wildfire.The press didn’t just spread truth. It spread trouble.

Five hundred years later, we’re at another hinge point. The printing press made ideas replicable.Now, generative AI makes them recombinable.And once again, the creators are being left behind.

The New Scribes Are Machines

Large Language Models — ChatGPT, Claude, Gemini and others — are devourers. They ingest everything: books, blogs, lyrics, academic papers, forums. From that vast feast of language they generate new text — fluent, plausible, synthetic.

These systems are trained on human creativity, yet respond as if no human ever existed.The reader types a question.The machine answers.The author is gone.

When Fair Use Meets Fuzzy Ethics

The legal ground is shifting fast.

• The New York Times is suing OpenAI.

• Universal Music Group is suing Anthropic.

• Meta faces accusations of training on pirated books from LibGen.

At the centre of it all is one phrase with sharp teeth: “irreparable harm.”

Are AI companies committing theft when they train on your life’s work?Some courts say no — if no sentence is copied, no harm is done.But the creative economy runs on recognition, not just revenue.

Fair Use vs Fair Dealing: A 30-Second Guide

Context	Principle	Status
US – Fair Use	Flexible, case-by-case: purpose, nature, amount, market effect.	Decided by precedent; active litigation ongoing.
UK/EU – Fair Dealing & Text-and-Data Mining	Specific exceptions with opt-outs.	Permitted for research, not commercial training.
Bottom line: “Legal” isn’t the same as “fair.” Creators deserve attribution and licensing, even when law lags behind ethics.

The Real Risk Isn’t Copying - It’s Replacement

No one’s selling photocopies of our books. The harm is subtler:a student asks, “Explain the French Revolution,” and receives a fluent composite of a dozen authors — mine among them — without a single name attached.

When a librarian once emailed to ask who wrote a paragraph she loved, there was no author to point to. That’s not citation; that’s substitution.

For those of us who write, illustrate, and design — especially for children, where clarity and originality matter — this isn’t theft. It’s erasure.

The Historical Habit of Erasure

The disappearance of creators isn’t new.From medieval illuminators to anonymous studio artists, from ghostwriters to unpaid interns, countless names have been scrubbed from the record. The difference now is scale.AI doesn’t forget some of us. It forgets everyone.

What Can Be Done

It’s tempting to sigh and let the tide rise. But the press didn’t stop the monks from writing, and this won’t stop us either.

Five Moves for Humans Who Make Books

1. Contract for provenance: Insert clauses that ban unlicensed text-and-data mining on manuscripts.

2. Mark your work: Maintain visible credits, embedded metadata, and consistent author names (ISNI, VIAF, ORCID).

3. Metadata moat: Keep bylines identical across platforms; discoverability is your defence.

4. Use AI on your terms: Draft schedules, comp sets, blurbs — yes. Replace your voice - no.

5. Join collective action: Support transparent, remunerated licensing frameworks. Visibility is legacy.

The Pressure of Progress

Gutenberg changed the world with rag pulp, movable type, and pressure. What he made was messy, glorious — and dangerous.We’re at the edge of a similar moment. But this time, the raw material isn’t cloth. It’s us.

Let’s not be turned to pulp without protest.

The Next Chapter: Standing on Silicon Shoulders

Yet let’s not be turned to pulp without purpose either.

Every creative leap once looked like heresy.Photography was a “mechanical cheat.”Desktop publishing was “the death of design.”Each time, the tools matured — and so did we.

AI isn’t the enemy. It’s a blunt, imperfect, immensely powerful instrument.It needs boundaries, not banishment.Yes, creators should be paid when their work trains a model. Yes, transparency matters. But let’s not throw the creative baby out with the synthetic bathwater.

AI can save time. It can help a dyslexic child write their first poem, or give a lone illustrator layout skills once locked behind InDesign.It can empower a single creator to do what once took a whole house.

This isn’t the death of creativity. It’s a reshaping.

The next generation of creators won’t just be painters, writers, or editors — they’ll be conductors of tools: fluent in empathy, imagination, and code.

Progress is pressure. We survive it by keeping our fingerprints on the page.

A Short History of Technological Panics in Publishing

Or, How We Learned to Stop Worrying and Love the Machine

1470s - Printing PressPanic: “This will destroy beautiful handwriting.”Reality: Bibles, broadsheets, bawdy poems by the cartload.Moral: Every age thinks its pen is the purest.

1814 - Steam PressPanic: “Speed will ruin quality.”Reality: It did — and readers loved it. Cheap print made literacy mainstream.Moral: Democratization always looks like degradation to the guild.

1880s - Linotype and Hot MetalPanic: “Machines will replace craftsmen.”Reality: They did, but newspapers boomed.Moral: Convenience wins — nostalgia follows.

1960s - Offset LithographyPanic: “The end of real printing.”Reality: Print became cleaner, faster, cheaper.Moral: Better methods never feel romantic until they’re obsolete.

1980s - Desktop Publishing & QuarkXPressPanic: “Real designers use scalpels!”Reality: They used Quark by Monday.Moral: Once you’ve kerned by mouse, you never go back.

1990s - The InternetPanic: “No one will read books again.”Reality: Everyone read more - they just stopped paying.Moral: Access doesn’t kill appetite; it changes the menu.

2007 - Kindle & EbooksPanic: “Print is dead.”Reality: Print recovered; ebooks found their niche.Moral: Story beats format.

2020s - Generative AIPanic: “This will erase human creativity.”Reality (so far): It’s redefining what ‘human’ means.Moral: The tools change. Taste endures.

Progress applies pressure. We’ve survived hot metal, Quark, and the Kindle because we kept our names on the work. Keep the names. Fix the licences. Use the tools. Don’t let the fibres forget who they were.