Part 2 of the ‘Quiver, Don’t Quake’ Book of the Year Series
Missed Part 1? For an introduction, check out this post:
The Industry That Forgot What Creativity Actually Is
Here’s an uncomfortable truth about publishing: an industry built entirely on human creativity has surprisingly little systematic understanding of how creativity actually works.
We can spot it when we see it. We know a brilliant manuscript from a mediocre one. We can articulate (sometimes) what makes prose sing or a narrative compel. But ask most publishing professionals to explain the psychology of creativity – the actual cognitive and emotional processes that generate original work – and you’ll get literary metaphors, not frameworks.
We speak of “the muse,” of “inspiration striking,” of writers who are “naturally gifted.” We talk about creativity as if it’s lightning – mysterious, unpredictable, bestowed by forces beyond understanding. And yeah, it makes us feel good. Important, even.
This wouldn’t matter, except that our fuzzy understanding of creativity is now colliding with a technology that demands precision. When we can’t articulate what makes human creativity unique, we can’t defend it. When we don’t understand the mechanics of the creative process, we can’t intelligently integrate new tools into it.
This is where Nadim Sadek’s Quiver, Don’t Quake makes perhaps its most valuable contribution. Before discussing AI at all, Sadek spends his first chapter doing something publishing desperately needs: establishing a rigorous psychological framework for understanding human creativity.
And in doing so, he reveals why the publishing industry’s response to AI has been so confused.
What Creators Say About Creativity
Sadek begins not with theory but with voices – asking creators themselves to describe the experience of creation. The results are striking in their consistency across cultures and disciplines.
Yayoi Kusama (Artist): “My art originates from hallucinations I have been seeing since my childhood. I translate the hallucinations and obsessional images that plague me into sculptures and paintings.”
Steve Jobs (Co-founder of Apple): “Creativity is just connecting things. When you ask creative people how they did something, they feel a little guilty because they didn’t really do it, they just saw something.”
Martha Graham (Dancer & Choreographer): “There is a vitality, a life force, an energy, a quickening that is translated through you into action, and because there is only one of you in all of time, this expression is unique.”
Pablo Picasso: “Every child is an artist. The problem is how to remain an artist once he grows up.”
Four distinct voices, four different creative fields – yet the same themes emerge:
- Creativity is translation – turning inner experience into shareable form
- Creativity is connection – seeing relationships others miss
- Creativity is innate – we’re born with it; the challenge is preservation
- Creativity requires courage – willingness to step into uncertainty
For publishers, this should be revelatory. Creativity isn’t a rare gift bestowed on a talented few. It’s a universal human capacity.
The grandmother in Lagos who’s never written a book isn’t “not creative.” She simply hasn’t had access to the tools (including time, education, and craft mastery) to translate her inner world into publishable form.
This reframing matters enormously for understanding AI’s role. If creativity is universal but expression is constrained by craft barriers, then tools that lower those barriers don’t threaten creativity – they liberate it.
Which with bitter irony is fuel to the Luddite Fringe fire. They don’t want anyone else being enabled, by AI or otherwise. AI is not the threat to them. It’s AI enabling competition that has them screaming that the sky is falling.
What Psychologists Say: The Science of Creation
From artists’ testimonies, Sadek turns to psychological research – and here the framework becomes even more useful for publishers.
The Dual-Process Model
The key insight comes from psychologists Scott Barry Kaufman and Daniel Kahneman, who describe creativity as a “dance” between two distinct cognitive modes:
System 1: Fast, Intuitive, Feeling-Driven
- Automatic and associative
- Operates below conscious awareness
- Generates hunches, gut feelings, sudden connections
- The “aha!” moment, the flash of inspiration
- What we typically call “the creative spark”
System 2: Slow, Analytical, Deliberate
- Conscious and effortful
- Logical and structured
- Evaluates, refines, organises
- The editing mind, the critical faculty
- What we might call “the craft”
Here’s what publishers typically misunderstand: We tend to romanticise System 1 (the spark, the muse, the inspiration) whilst undervaluing System 2 (the hard work, the revision, the structure). We speak reverently of “raw talent” whilst (unless we are editors ourselves) treating editorial refinement as mere “polishing.”
But as Sadek demonstrates, both systems are essential – and creativity is precisely the fluid movement between them.
Consider the novelist:
- System 1 generates the initial idea: “What if a character could only tell the truth?”
- System 2 asks: “How would that work? What consequences? What structure?”
- System 1 imagines a scene: the character at a job interview, unable to lie
- System 2 writes the scene, checks consistency, revises for clarity
- System 1 feels something’s wrong, the scene is flat
- System 2 analyses why, restructures
- Back and forth, umpteenth times, until the book is finished
The creative act isn’t the spark alone. It’s Kaufman and Kahneman’s dance.
Where Publishers Go Wrong: Real-World Scenarios
Armed with this framework, we can identify precisely where publishing’s thinking about AI goes awry. Let me illustrate with fictional but realistic scenarios any publishing professional will recognise:
Scenario 1: The Rejection Letter Dilemma
The Situation: An editor or literary agent receives a memoir submission. The story is extraordinary – a refugee’s journey that offers profound insights into resilience and identity. The voice is authentic, raw, compelling. But the manuscript has structural problems: chapters are uneven, chronology jumps confusingly, some scenes are underdeveloped whilst others drag.
Traditional Response: Rejection. Period. Occasionally, rejection with feedback: “This needs substantial structural work. Please revise and resubmit.”
The Problem: The author has a powerful System 1 insight – lived experience that matters. But they lack System 2 craft mastery to structure it effectively. They might spend another two years trying to learn narrative architecture, and in doing so focus so much on structure that they weaken the voice that matters, because they were not told that was where the value lay. Or more likely still, they give up entirely.
The AI-Assisted Reality: The same author uses AI to explore different structural approaches. They try chronological versus thematic organisation. They test where flashbacks work best. Here’s the thing: The AI doesn’t write the memoir – it helps the author find the structure that best serves their authentic voice.
Three months later (likely much sooner), the manuscript returns. The story is the same. The voice is the same. But the structure now serves the content. Of course, had the author used AI in the first place there would have been no re-submission.
Except… Maybe the author didn’t dare use AI in the first place because of so much irrational hostility out there in writing-land, and that book never got written at all.
What the Editor / Agent / Publisher / Fellow Author Should Recognise: This isn’t AI writing. It’s AI helping the author develop their System 2 craft whilst preserving their System 1 authenticity. The rejection should have been “fascinating voice, needs structural support” – and now that support is accessible.
And here’s the exciting part: the editor or agent can tell the author to go do it themselves, or the agent or editor could themselves use AI to guide the author.
Publishing’s Challenge: Learning to evaluate System 1 authenticity separately from System 2 execution. The former is what we’re actually publishing; the latter can increasingly be assisted.
Scenario 2: The Slush Pile Paradox
A reminder from Part 1: neither Nadim Sadek nor myself routinely use the derogatory term “slush pile”, but for this audience the term is an industry label I guess I need to deploy here.
The Situation: A small independent publisher reviews their slush pile statistics. Five years ago, they received 2,000 unsolicited submissions annually. This year: 8,000. The editorial assistant is overwhelmed.
Traditional Response: Panic. “AI is flooding us with garbage. We need better filters.”
What’s Actually Happening: Many of those 6,000 additional submissions aren’t “AI-generated books.” They’re from writers who previously would never have submitted because they knew their craft wasn’t publishable-quality. AI has helped them develop System 2 competence, so now they’re taking the risk.
The Real Question: How many of those 6,000 have genuine System 1 insights worth publishing?
One Publisher’s Solution: Rather than filtering out “AI-assisted” manuscripts (impossible to detect reliably anyway), they redesigned their submission process around what Sadek would call System 1 evaluation:
- New Question 1: “In 200 words, explain why you specifically needed to write this book. What lived experience or insight do only you possess?”
- New Question 2: “What will readers feel when they finish your book? Not what will they learn or think – what will they feel?”
- New Question 3: “What makes your voice distinctive? If we read a paragraph anonymously, what would mark it as unmistakably yours?”
The Result: Submissions dropped 40% (writers without authentic System 1 insights self-selected out), but acceptance rates from the remaining pool increased dramatically. The publisher wasn’t filtering for craft anymore – they were filtering for creativity.
The Lesson: AI assistance with System 2 craft is making System 1 authenticity more visible, not less. Publishers just need to adjust what they’re looking for.
On a personal note, I’m reminded of a time back in nineteen-bow-and-arrow when desktop word processors shunted manual typewriters into the dustbin of history.
As a novice journalist I was assigned to write a post about the slush pile, and was assigned to a major London agency to spend a day in, back when my own novel-writing days were still in their infancy
Email submissions were still unheard of, or at least not an option, and I witnessed a ton of brown envelopes full of mostly unsolicited manuscripts land in reception, and I watched an intern open and evaluate them, often single glance at the covering letter landing a package on the Return To Sender table.
I was horrified! All that hard work! That could have been one of my unedited half-baked masterpieces! Not even looked at!
No need to waste time looking at the manuscript. I understood the rejection.
Then I read – or rather tried to read – the covering letter. No need to waste time looking at the manuscript. I understood the rejection. Busy agency. Hundreds of unsolicited submissions. Only 24 hours in the day.
But what a senior agent explained to me was that this “new-fangled” word processing nonsense was to blame. Now everyone thinks they can write the next bestseller. I blame Bill Gates. Submissions have quadrupled since MS Word came along. Oh for the days when authors had to spend hours changing typewriter ribbons and correcting typos with liquid paper. The sky is falling!
Then a younger agent took me to one side and confided, “When he’s not around I skim over the submissions in case he’s rejected any hidden gems. I was crap at spelling at school, but loved writing stories. My teacher encouraged me to focus on the story, not the presentation. Now I’m a lit-agent!”
Scenario 3: The Developmental Edit Crisis
The Situation: A mid-sized publisher with a strong editorial team notices their developmental editors are spending 60% of their time on what they call “mechanical fixes” – pointing out structural inconsistencies, suggesting transitions, identifying repetitive passages, flagging POV slips.
The Editorial Director’s Worry: “If AI can handle mechanical fixes, what’s left for our editors to do?”
What Happened: They ran an experiment. For three months, they asked authors to use AI to address mechanical issues before submitting revised manuscripts. Authors were given specific prompts: “Analyse this manuscript for structural inconsistencies,” “Identify repetitive language,” “Check POV consistency.”
The Results Were Surprising:
- Time saved: Developmental editors reduced mechanical feedback time by 70%
- Time reallocated: That saved time went to deeper questions: “Your protagonist makes this choice in chapter 12, but given their background in chapter 3, would they actually do this?” “The emotional arc peaks too early. The reader has nowhere to go after chapter 8.” “Your theme of forgiveness is undermined by how you resolve the antagonist’s storyline.”
Author Feedback: “The AI fixed my technical problems, but my editor helped me understand what my book was really about.”
The Revelation: When System 2 mechanical work is handled by AI, editors can focus on what Sadek would call the “creative psychology” – helping authors understand their own System 1 insights more deeply.
One Editor’s Reflection: “I used to spend 40% of my time saying ‘this scene doesn’t work’ and explaining why structurally. Now I spend that time asking ‘what is this scene trying to make the reader feel?’ and helping the author articulate that. It’s more human editing, not less.”
Scenario 4: The Copy-Editor’s Evolution
The Situation: A publisher employing twelve freelance copy-editors received a panicked email from one of their most experienced editors: “I just edited a manuscript that was clearly AI-assisted. It was grammatically perfect, stylistically consistent, properly formatted. I made almost no technical corrections. Am I becoming obsolete?”
The Publisher’s Response: They organised a workshop exploring not what copy-editors do, but why it matters.
The Realisation: Copy-editing has always been about two distinct functions:
- Technical correction (System 2): grammar, consistency, formatting
- Voice preservation (System 1): ensuring editorial changes don’t homogenise the author’s distinctive style
When AI handles function 1, function 2 becomes essential.
New Copy-Editor Brief: “Your job isn’t to catch typos anymore – AI does that. Your job is to be the guardian of the author’s voice. If AI suggestions have made the prose too ‘perfect,’ too generic, restore the quirks that make this author distinctive.”
Practical Example: An AI might correctly change “she staggered, lurched, careened down the street” to “she staggered down the street” for concision. But a good copy-editor might restore it, recognising that the piled-up verbs capture the character’s drunken perspective in a way the “correct” version doesn’t because an AI had never been drunk and doesn’t get the author’s choice of words.
One Copy-Editor’s Reflection: “I thought AI would replace me. Instead, it’s freed me to do the work I always found most satisfying – protecting what makes each author’s voice special. I’m less proofreader, more voice consultant.”
The View From The Beach: Yeah, proofreaders probably are going to struggle to find work. But hey, now’s the time to upskill!
Scenario 5: The Commissioning Conundrum
The Situation: A commissioning editor at a major trade publisher is evaluating two proposals for books about climate anxiety:
Proposal A: From an established environmental journalist. Perfect proposal – clear structure, compelling case studies, authoritative voice. Impeccably written. The author mentions using AI to “streamline the proposal process.”
Proposal B: From an unknown climate scientist. Decent proposal – solid research, some structural issues, occasional grammatical errors. Raw but passionate. Clearly written without AI assistance.
Traditional Logic: Commission Proposal A. It’s professional, polished, from a proven author.
What the Editor Actually Did: She commissioned Proposal B.
Her Reasoning (Using Sadek’s Framework):
Proposal A showed strong System 2 execution: Clear structure, good writing, professional presentation. But she couldn’t identify what unique System 1 insight the journalist brought beyond “I’m good at explaining complex topics.”
Proposal B showed weaker System 2 execution but powerful System 1 authenticity: The scientist had spent twenty years studying ice core samples in Antarctica. Her proposal contained this paragraph:
“When you drill two kilometres into ancient ice, you’re holding in your gloved hands the atmosphere breathed by humans who built the pyramids. You can see their world in the trapped air bubbles—cleaner, cooler, more stable. And you know, with absolute certainty, what we’re losing. That knowledge sits in your chest like its own kind of frozen core. This book is about learning to live with that knowledge without being paralysed by it.”
The Editor’s Reflection: “Proposal A was technically superior. AI helped that author optimise structure and language. But Proposal B had something no AI could provide or enhance – authenticity born from two decades of lived experience. My job isn’t to commission the most polished proposals. It’s to identify voices that matter.”
The Commissioning Shift: From “who writes best?” to “who has something irreplaceable to say?”
Six months later: The scientist’s manuscript arrived. Structurally rough, as expected. The publisher paired her with a structural editor who (with AI assistance) helped organise the material whilst preserving that distinctive voice. The book became an award-winner – specifically praised for its “unique insider perspective that transcends typical climate journalism.”
The Lesson: When System 2 craft becomes accessible to everyone, publishers who focus on System 1 authenticity will discover voices previously excluded by craft barriers.
The Flow State: What We’re Really Looking For
One of Sadek’s most useful contributions is highlighting psychologist Mihaly Csikszentmihalyi’s concept of “flow” – that magical state where creation feels effortless, where hours pass unnoticed, where the work pours out naturally.
It’s reaching for coffee only to find it’s gone ice-cold because you’ve been writing for four hours straight without noticing.
Every writer knows this experience, though they might not know the name. It’s reaching for coffee only to find it’s gone ice-cold because you’ve been writing for four hours straight without noticing. It’s when the words flow faster than you can type them, when you’re not consciously choosing each sentence but channelling something that feels larger than yourself.
Csikszentmihalyi describes flow as the perfect synchronisation of System 1 and System 2. The intuitive and analytical minds work in seamless harmony. No internal conflict, no self-doubt, just pure creative engagement.
But here’s what’s crucial: Flow requires the right balance of challenge and skill. Too easy, and we’re bored. Too difficult, and we’re anxious. Flow exists in that sweet spot where we’re stretched but not overwhelmed.
For many aspiring creators, the craft requirements of traditional publishing prevent flow. They’re so overwhelmed by technical demands (How do I structure this? Is my grammar correct? Does this scene work?) that they never reach the state where their authentic voice can emerge.
A Publisher’s Flow Story
Okay, so another scenario. A publisher specialising in translated literature describes this phenomenon:
We had a brilliant Albanian writer whose work we desperately wanted to publish. Her stories – drawn from her childhood during the communist era – were extraordinary. Unique perspective, powerful emotional truth, that System 1 authenticity Sadek talks about.
But her English wasn’t strong enough for publication, and professional translation was prohibitively expensive for our small press. We were stuck.
Then she started using AI translation as a first pass – not as final product, but to get 80% of the way there. She’d translate her Albanian text through AI, then work with us to refine it, preserving her voice whilst improving clarity.
She told me: ‘For the first time, I could focus on whether the English captured my meaning, not whether my grammar was correct. I could be in flow with my stories again, not anxious about language mechanics.’
The result was published last year and longlisted for an international prize. The judges specifically praised its ‘distinctive voice that feels authentically personal despite being translated.’
That’s Sadek’s argument proved: AI handling System 2 technical work enabled her System 1 creativity to shine.
The Publisher’s Role in Enabling Flow
This suggests a profound shift in how publishers think about their function:
Old model: We evaluate the finished product – is this manuscript publishable?
New model: We help authors reach the creative flow state where their best work emerges – what support do they need?
That support might include:
- AI tools for System 2 craft work
- Editorial guidance on System 1 insight development
- Collaborative partnerships where technical execution is assisted
- Focus on voice preservation over technical perfection
One Editorial Director’s Reflection: “We used to be gatekeepers asking ‘is this good enough?’ Now we’re midwives asking ‘what does this author need to deliver their best work?’ Sometimes that’s AI assistance. Sometimes it’s developmental editing. Sometimes it’s just permission to be weird and authentic rather than conventionally ‘good.'”
The Failure Loop: What Goes Wrong in Practice
Sadek is honest about the risks – what he calls the “Failure Loop.” Here’s how publishers are seeing these failures manifest (again, fictional scenarios derived from real-life feedback):
Case Study: The Over-Reliant Author
The Situation: An editor received a manuscript from a debut novelist. Technically flawless – perfect grammar, consistent POV, well-structured scenes. But something felt off.
The Problem Revealed: In their second meeting, the author admitted: “I fed my rough draft to AI and asked it to ‘improve the prose.’ It did such a good job, I used most of its suggestions.”
What Got Lost: The author’s original draft had a distinctive, slightly fragmented style that mirrored the protagonist’s anxious mental state. The AI “corrected” this into smooth, conventional prose – technically superior but emotionally flat.
The Resolution: The editor retrieved the original draft. Together, they identified which “AI improvements” to keep (genuine errors) and which to reject (stylistic quirks that were actually artistic choices).
The Author’s Reflection: “I over-trusted the AI. I assumed ‘better grammar’ meant ‘better writing.’ It took my editor to show me that my ‘mistakes’ were sometimes my voice.”
For Publishers: This is the new editorial skill – helping authors understand when AI assistance enhances their work versus when it homogenises it.
Case Study: The Generic Pull
The Situation: A small press noticed a troubling pattern. Over eighteen months, they’d published twelve debut novels – six AI-assisted, six traditionally written. Reader reviews for the AI-assisted books consistently used words like “competent,” “professional,” “well-crafted” – but rarely “moving,” “unforgettable,” or “unique.”
The Investigation: The editorial team re-read all twelve manuscripts, specifically looking for what Sadek calls “the quirks that make writing distinctive” – unexpected metaphors, unconventional structures, stylistic choices that broke rules.
The Finding: The AI-assisted manuscripts had significantly fewer of these quirks. The authors had used AI to smooth out rough edges, but in doing so, had sanded away distinctiveness.
One Example: An author’s original draft described grief as “like being underwater but still breathing, everything muffled and wrong.” AI suggestions kept steering toward more conventional metaphors: “like a heavy weight,” “like walking through fog.” The author, trusting AI’s statistical probability, chose the conventional option.
The Publisher’s New Policy: Authors using AI assistance must also submit a “voice sample” – a chapter from their rough draft showing their natural, unassisted style. Editors use this to ensure AI hasn’t eroded distinctiveness.
One Author’s Grateful Response: “Seeing my rough draft next to my AI-polished version was revelatory. I’d let the AI turn my voice into everyone’s voice. Now I use it for grammar and structure, but I protect my weird metaphors and unconventional choices. That’s what makes my writing mine.”
Case Study: Losing the “Why”
The Situation: A literary agent received a submission for a climate fiction novel. The query letter was perfect – crisp premise, clear stakes, professional presentation. But when she requested the full manuscript, something felt hollow.
The Problem: In their call, the author struggled to articulate why this book mattered to them personally. They’d used AI to develop the premise, structure the plot, even generate character backstories. The result was technically accomplished but emotionally empty.
The Agent’s Question: “Why did you write this book? What made you passionate enough about climate fiction to spend a year writing it?”
The Author’s Honest Answer: “I asked AI what kind of books were selling in the climate fiction market. It suggested this premise. I thought if I could execute it well, I’d get published.”
What Was Missing: Any System 1 authenticity – no personal connection to the material, no lived experience informing the narrative, no emotional stake in the themes.
The Agent’s Advice: “Don’t ask AI what books the market wants. Ask yourself what book only you can write. Then use AI to help you write it well.”
For Publishers: The crucial question in evaluating AI-assisted work becomes: “Does this author know why they’re writing this? Is there authentic intent, or are they just producing content the algorithm suggested?”
What This Means for Publishing’s Editorial Function
If we accept Sadek’s psychological framework, the implications for publishing are profound:
The Gatekeeping Shift
Scenario: The Two-Manuscript Test
Imagine an editor receives two manuscripts on the same morning:
Manuscript A: Memoir from an established author. Beautifully written, perfectly structured, emotionally intelligent. The author acknowledges using AI for “structural guidance and consistency checking.”
Manuscript B: Memoir from an unknown author. Decent writing with some rough patches, occasionally meandering structure, but a voice so distinctive and honest it makes you catch your breath. Written entirely without AI assistance.
Traditional publishing logic: Commission A – it’s professional, polished, from a proven author.
Sadek’s framework suggests: Ask different questions:
- Which author has irreplaceable System 1 insight?
- Which voice is truly distinctive?
- Which manuscript teaches readers something they couldn’t learn elsewhere?
- Where is the emotional authenticity strongest?
In this case: Both might be publishable, but for different reasons. Manuscript A shows how AI can enhance an already strong author’s System 2 execution. Manuscript B shows raw System 1 authenticity that matters despite craft roughness.
The New Gatekeeping:
Old question: “Is this manuscript professionally executed?” New question: “Does this author have something authentic to say, and is their voice distinctive enough to matter?”
Old rejection: “Your grammar and structure need work.” New rejection: “I don’t understand what makes your perspective unique, or why readers need this book specifically from you.”
This is harder work – but more honest work. It’s what publishing should have been doing all along.
The Editorial Evolution: Real Skills for Real Challenges
Different editorial roles are evolving in specific, practical ways:
Copy-Editors Becoming Voice Consultants
New Skill Set:
- Recognising when AI suggestions have homogenised an author’s distinctive style
- Identifying which “errors” are actually artistic choices (sentence fragments for effect, unconventional punctuation that creates rhythm)
- Restoring quirks that make a voice memorable
Practical Example: A thriller writer uses short, punchy sentences. AI smooths these into standard compound sentences for “flow.” The copy-editor recognises the original staccato rhythm matched the genre’s tension and restores it.
Developmental Editors Becoming Creative Psychologists
New Skill Set:
- Helping authors articulate what they’re really trying to create
- Identifying when System 2 craft is masking weak System 1 insight
- Asking questions that excavate deeper authenticity
Practical Example: Author presents a “romance about second chances.” Through developmental questioning, editor discovers it’s actually about forgiving yourself for past versions of yourself – a much more interesting emotional core. The romance becomes vehicle, not premise.
Commissioning Editors Becoming Cultural Arbiters
New Skill Set:
- Identifying which voices represent genuinely new perspectives versus which are technically competent but familiar
- Understanding cultural moments and which stories address them authentically
- Evaluating whether AI assistance has enhanced or eroded an author’s distinctive contribution
Practical Example: Two proposals about pandemic experiences arrive. One is polished, AI-structured, hits all expected beats. The other is rougher but written by an ICU nurse whose lived experience offers insights no amount of research could provide. The commissioning editor chooses authentic perspective over technical polish.
The Psychological Trap Publishers Must Avoid
Sadek highlights a psychological phenomenon highly relevant to publishing: the tension between algorithm aversion (dismissing AI recommendations because they challenge our intuition) and automation bias (over-trusting AI because it seems authoritative).
Scenario: The Title Debate
The Situation: A debut author presents their novel with the working title “The Weight of Memory.” AI analysis of comparable titles, genre conventions, and reader data suggests “What We Carried” would perform 40% better commercially.
Algorithm Aversion Response: “That’s a data-driven suggestion. It lacks soul. We’re trusting human instinct here.”
Automation Bias Response: “The AI is probably right. Let’s change it.”
Calibrated Trust Response: The editor considers:
- What does the original title mean to the author? (System 1 authenticity)
- Is there a thematic reason it matters? (Artistic integrity)
- Does AI data reveal a genuine market insight? (Commercial reality)
- Could a compromise honour both? (Creative solution)
The Resolution: Through discussion, they discover “What We Carried” actually better captures the novel’s theme of inherited trauma. The author was attached to “Weight” but admits “Carried” is more precise. The AI data revealed commercial truth; the human conversation revealed artistic truth. Both mattered.
The Skill: Knowing when to trust data and when to trust intuition – and recognising these aren’t always opposed.
Creativity as Universal: Practical Implications
Sadek’s foundational belief – that creativity is innate to being human – has profound practical implications for publishers:
The Manuscript That Shouldn’t Work (But Does)
A Publisher’s Story:
“We received a submission that broke every rule we teach in creative writing workshops. Inconsistent tense, meandering structure, occasional grammatical quirks that made us wince. By traditional standards, it wasn’t publishable.
But the voice – my God, the voice. It was a Zimbabwean grandmother’s memoir of surviving Mugabe’s regime, written in English that occasionally slipped into Shona syntax and rhythm. Some sentences were fragments. Others ran on like oral storytelling.
The author apologised in her cover letter: ‘I know my English isn’t good. I used AI to help, but it kept changing my voice to something that didn’t sound like me, so I used it minimally.’
We commissioned it. It won awards. Reviewers praised its ‘distinctive voice that feels authentically oral despite being written.’
That’s what Sadek means by System 1 authenticity mattering more than System 2 polish. Her lived experience – that spark of genuine insight – was irreplaceable. The craft could have been assisted more, but not at the cost of her voice.
And here’s the thing: ten years ago, she might never have submitted at all, assuming her English wasn’t ‘good enough.’ AI gave her just enough confidence to try – even though she wisely didn’t let it homogenise her voice.”
Publishers who deeply understand the psychology of creativity will thrive in the AI age
Psychology as Publisher’s Competitive Advantage
The irony is exquisite to the point of being delicious: publishers who deeply understand the psychology of creativity will thrive in the AI age precisely because AI cannot replicate that understanding.
AI can analyse. It can execute. It can optimise.
But it cannot feel. It cannot intuit. It cannot make the System 1 leaps that identify which stories matter.
Sadek’s framework gives publishers a language for what they’ve always done implicitly:
- Recognising the System 1 spark that makes a manuscript special
- Understanding the System 2 craft that can be taught (or AI-assisted)
- Identifying when authors are in creative flow versus when they’re producing by rote
- Spotting the difference between authentic voice and generic output
The publishers who thrive won’t be those who resist AI. They’ll be those who understand human creativity deeply enough to:
- Leverage AI for System 2 work whilst preserving System 1 authenticity
- Recognise when AI has been used well versus poorly
- Identify genuinely novel voices in a sea of technically competent content
The Practical Checklist for Evaluating AI-Assisted Manuscripts
When reviewing manuscripts where AI assistance is acknowledged (or suspected), ask:
System 1 Questions (Authenticity):
- Can the author articulate why they specifically needed to write this?
- Is there lived experience here that feels irreplaceable?
- Does the voice have distinctive quirks, or is it generically “good”?
- What will readers feel (not think, not learn – feel)?
System 2 Questions (Execution):
- Has AI assistance enhanced clarity without eroding distinctiveness?
- Are there signs of over-reliance (prose that’s too perfect, too conventional)?
- Has the author maintained agency over artistic choices?
- Can the author explain their structural and stylistic decisions?
The Integration Question:
- Has AI helped this author reach creative flow, or has it created distance from their material?
This is what Sadek means when he says AI makes us “more human.” By handling the analytical tasks, it forces us to lean on our uniquely human capabilities: intuition, taste, cultural fluency, emotional resonance.
These have always been publishing’s real competitive advantages. AI just makes them non-negotiable.
And for those eight billion creative sparks Sadek celebrates – those voices previously silenced by craft barriers – AI may finally lower the wall enough for their System 1 authenticity to shine through.
The question for publishing isn’t whether to engage with this reality. It’s whether we’re ready to judge manuscripts on what’s always mattered most: whether they have something irreplaceable to say, and whether they say it in a voice that’s unmistakably, authentically human.
In the next article in this series, I’ll review and expand on what Nadim Sadek says architecture, music, medicine, and other creative industries have learned about AI integration – lessons publishing ignores at its peril.
But start here. Start by understanding what creativity actually is.
Because only then can you understand what role AI should – and shouldn’t – play in it.
This post first appeared in the TNPS LinkedIn newsletter.
