Nadim Sadek is Founder and CEO of Shimmr AI, which produces automated advertising that sells books. He’s also the author of a book, ‘Shimmer, don’t Shake - How Publishing can Embrace AI’, is a regular columnist for The Bookseller, and is preparing a collection of children’s stories for publication. Trained as a psychologist, his career has spanned the founding and growing of businesses in market research, whiskey, food, music and AI-brand management. He also reviews motorcycles on Boss Bikes Club YouTube channel, and manages a Warner-signed artist, Shaefri.
Your incredible career spans different industries and innovations. What inspired you to focus on publishing?
Don't raise expectations too high. I mean, it's been a fun career and a multi-faceted one, and I've enjoyed it. What inspired me to get involved with publishing is one of the first things that got me interested in careers.
When I was at school, I wrote a lot of stories. I tried to win the English prize for essays and stuff, and I was always very hopeful that I'd become a writer. But for whatever reason, my career took a more corporate path than that. I've returned to writing more recently. So, it's kind of a personal interest in the world of literature and books.
The thing that I suppose has characterised my career, without wanting to be self-aggrandising or congratulatory about it, is that I'm sort of a pattern breaker. Whenever there's a status quo, I have found myself changing the way it works in some way.
Around publishing, there has been a very long challenge of discoverability, and being able to crack the code of discoverability and change the way that works has drawn me to it. So, maybe it's just a marriage of that very early love of books and writing, and a business opportunity, that we saw.
How did your experiences in your early career shape the way that you approach your work today?
It’s what I described, breaking that mould. It's not quite as grand as that, but finding different ways to do things, sometimes more efficient, sometimes more effective, sometimes novel.
I did that through a market research career where I helped brands to find innovative ways to present themselves and to connect with the consumers. I then started several brands of my own in whiskey, music, food and indeed, published a book. Those were all slightly different.
The whiskey came in a different shaped bottle that always stood out, because it was a bit skew. The idea was to “live life at a tilt,” and the bottles were little signals that our instinct, and my instinct in particular, is to try and find a new angle on things. That's the way we're approaching publishing, too.
Your book “Shimmer, don't Shake” discusses the integration of AI and publishing. How do you see AI transforming the publishing landscape over the next few years?
The way I'm seeing it is that we're on a journey of three E's, so we're currently very focused on employing AI to increase corporate Efficiency. That's the first E. How can we make rights work more efficiently? Can we qualify manuscripts more fairly? Can we make sure that our sustainable footprint is improved? You know, all of these things AI can assist, are all to do with corporate efficiency.
The place we're going to for the next E is creative Emancipation. Not the writing of lots of books by AI, that's just boring. AI is made to be anodyne and harmless and modest in the way that it talks, so whatever it writes is pretty boring. What I’m talking about is to liberate, perhaps new art forms.
I work with a lot of Gen Z people, and they're characterised, people say, by a lack of commitment to careers or to really believing in any particular corporate structures and so on. They get frustrated by this characterisation, because actually, Gen Z is full of creativity, but it’s quite hard to articulate it or to manifest it to the rest of society.
So can I see a Gen Zer mashing up, to use an old phrase, a video game with a cryptocurrency, with a bit of music, with a bit of literature, into a new art form as yet undefined. Can, and will, AI assist in that coming to fruition? Absolutely, you probably couldn't do it without AI.
Creative emancipation is a really exciting part of what AI will bring to the creative industry, and not just the encouragement of new art forms. I would contend that if there's 8 billion people alive on earth today, there are probably 8 billion creatives as well. It's just that they don't have means of expression, and if AI can somehow assist them in expressing that, again, not explicitly writing, but showing them ways in which they can articulate their thoughts, speak or harness other creative forms to make their feelings, their insights and their creative sparks visible, I think that's great.
If you're wondering where the third E has gone, we went from efficiency to emancipation with increasing Excitement.
Do you see AI revolutionising book sales?
Revolutionising is always a pretty big claim, but in some of the due diligence that we did as a business before entering publishing, it was said by many that there are about 100 million books published in English alone, never mind Mandarin, Portuguese, Arabic and all the others. Yet only about 5% of those are actively advertised.
Now, advertising isn't the be all and end all of anything, but it is a form of bringing things to people's attention. When you bring things to people's attention, you tend to lead to engagement, and engagement can lead to commercial interaction, which is a fancy term for they might buy it from you. So, if you make a book visible to somebody and it somehow connects with them, they might actually purchase it, and that's got great benefits. The publisher makes more of their catalogue, the author gets more reach and recognition, and the reader is matched to something that suits them better than blindly trying to find their way through 100 million books.
So, can AI assist in discoverability by matching works of art with audiences? Absolutely yes, it can. By understanding peoples’ interests and behaviours against the DNA of books, AI can revolutionise book sales just by better matching pieces of work to readers’ mindsets.
How can authors and publishers maintain the authenticity and creativity that defines great literature?
There are some fairly simple, concrete, rudimentary things to do. For example, if it's written by AI, have a note that says it's written by AI. If you've researched with AI and mentioned that you've employed a bit of AI to do some of your research, there's no harm in doing that.
I think we should all just be without shame in doing so. It's a helpful thing, whether you do it to gather information, or whether you do it as I sometimes do, to help you form better arguments. I write a regular column for The Bookseller, and every time I've completed an article, I put it into an AI and I say, contradict me. It tells me what an idiot I am, where my arguments are poor, where there are flaws, and it’s helpful. Being open about the fact that you're employing AI, collaborating with it, or having an AI companion, whatever way you want to frame the relationship, is helpful.
The other thing to remember is that, essentially, AI is a prediction machine. But if you think about it, one of the core definitions of creativity is unpredictability. When you read a great novel, it's that moment where you think, “Oh, my goodness, I would never have thought of that.” Or, a scenario is described and brought to life in your mind in a way that you never could have anticipated. That's creativity.
Even something like a historical narrative. It's the surfacing of an insight about why this military group overwhelmed another place, and what the morals of the thing were. It's so often about surfacing the unpredictable. So if you've got a machine which is defined by predictability, and you've got, on the other hand, what we humans do, which is essentially to be a bit wild and unpredictable, I don't see that AI is a great threat to the core creative outputs that great writers have.
Some critics worry that AI might make publishing too similar. How does Shimmr help ensure that diverse and unique voices are still heard?
It's actually an accidental byproduct of our whole approach, but what we're finding is that we're really enabling a lot of more muted voices to be much more present. For example, the LGBTQIA+ community is struggling to advertise its work on most of the major platforms, because the platforms have said you may not target the LGBTQIA+ community (this is actually for the protection of these audiences, but it has the opposite impact for authors trying to reach them with relevant works).
AI is actually able to help you connect in a less obvious way, by the sorts of things that you're interested in, rather than your identity. It's one way that we're helping to bring the mute a new voice.
The other thing is that we bring into the light things which are hidden – those which can't have a lot of human capital expended on them or can't have marketing campaigns in the traditional form invested in them. Those works tend to languish in the dark, and by having low cost, high impact ways of surfacing them with automated advertising, you can revivify some of those voices that have been lying quiet for too long. A lot of works which haven't gone anywhere for a while are being brought back into public attention by our automated advertising.
Traditional book promotion does have its limitations. How do you feel that AI driven strategies enhance the connection between books and their target audiences?
When I talk about AI, I'm usually thinking about the way that we are deploying AI, because it's what I know most about, and what we do is a reverse discoverability process. So typically, in almost every industry and also in publishing, you tend to think of, well, do we have a book for a 25 to 40 year old first-time mother who lives rurally rather than in an urban setting? And you define a few demographic and psychographic things, and you say, right, well, here's a book for her. We're not doing that. We're not doing the demographic part at all.
The book is our customer. We start by reading the book entirely and creating what we call BookDNA, which is a structural analysis of the book. There might be a story about mackerel fishing in Sweden. When I said to you “mackerel fishing in Sweden”, you had a picture of that, didn't you? We can’t help ourselves. We picture things.
But BookDNA is also a psychological analysis. If I tell you now that the book is also about melancholy, depression and cruelty, you're varying your pictures. They're getting a bit darker. They're getting a bit deeper, and if I say but also in that book, there are themes of family loyalty and ambition to leave Coastal Living and live in a big city, you’ve varied your pictures again. This book has become much more multifaceted in your mind, and that's what we do.
We analyse the BookDNA to find out all of the themes that are there, and then through psychologically themed advertising, we manifest those different aspects of the book that are going to be catchy emotionally and hook people. We know we're looking for people who might be interested in mackerel fishing, Sweden, melancholy, depression, cruelty, ambition, and coasts, and we go and find those sorts of people on the various platforms that aggregate audiences, whether it's Meta, Google, Pinterest, Amazon or something else.
We can match the book to the people that enjoy those themes, and that results in incredibly high rates of people engaging with the advertising and then going through a conversion page to say, yep, that's exactly what I was looking for. I'll buy it. So I think that's a change that AI has brought to book promotion. It's the reverse discoverability, where the book goes and finds its audience, rather than the other way around.
What ethical guidelines do you feel are necessary for the publishing industry to consider when integrating AI into their processes to ensure fairness and transparency?
I believe the really big picture is that we need to have a very robust and comprehensive system for recognising authorship and remunerating it. Sometimes that's called copyright, sometimes that's called rights, but I think about it as recognition and remuneration.
AI companies and all the models need content. They need more and more data to become better and better, and we want them to become better and better. We want them to be more robust, more dependable, to hallucinate, less, to be our companions, and to execute a lot of things in our lives. I believe having a proper recognition and remuneration system is what's principally needed.
There are some really interesting things happening. Obviously there are quite a few court cases going on about stolen copyright, which is obviously completely inappropriate, and that'll end up where it will end up, which I guess will be with payments of some sort.
I think outside of that more adversarial relationship, there's also licensing going on, one-to-one, AI model to particular publisher. I think that'll become more collective gradually, and we're probably going to have systems where every time an author produces a piece of work, they can say yes or no as to whether it could be employed in AI training. And then it just flows through contractually to payments to the publisher, payments to the author and the AI company gets higher quality content than it would otherwise.
It's quite difficult really. We should recognise that AI arrived with a bang, and it's a very young industry commercially. It wouldn't know how to have 10,000 conversations with 10,000 different publishers, so what we're doing now is facilitating the structures that have a much more collaborative approach to content and AI ingestion of really good work.
I would finish by saying that, though a bit crude and a bit silly, AI is what it eats, and we should feed it well. As long as it's recognised and rewarded properly, the more good AI can do, the better for us all.
