In this solo episode, I discuss the impact of converging technologies, Artificial Intelligence, Natural Language Generation (NLG) tools like GPT-3, and more on writing, authors, and the publishing industry.
My last AI show was in July 2019, 9 Ways That Artificial Intelligence (AI) Will Disrupt Authors and Publishing in the Next 10 Years, and although I’ve shared futurist segments on the show since then, I’ve been working on this as an updated view, and a look forward to the rest of the 2020s.
In this solo show, I discuss:
- Why this era of technological change, accelerated by the pandemic, is so important
- What is GPT-3 and why is it such a big deal?
- AI as co-creator: Embrace curiosity and play with artistic collaboration
- Impact on the commercial business model for writers, authors, and publishing
- What about books?
- Will AI tools make authors obsolete?
- “Data is the new oil,” the wealth generator of the digital age — and publishers have a lot of it. The publishing industry owns and controls the largest dataset of all.
- Why I’m so excited about the next decade of abundance — if we embrace the changes and opportunities ahead!
Check out previous episodes and resources on AI at www.TheCreativePenn.com/future
What started as preparation for this podcast episode actually turned into a book: Artificial Intelligence, Blockchain, and Virtual Worlds: The Impact of Converging Technologies On Authors and the Publishing Industry.
Available now as an ebook and coming soon in print and audio.
It covers:
- Writing in the age of AI, including Natural Language Generation models like GPT-3
- Copyright law, Blockchain for smart contracts, and micro-payments
- AI-assisted translation
- Voice technologies, streaming, and subscription
- Virtual worlds and augmented reality
- Global, digital, mobile. A wave of new writers.
In this episode, I cover Writing in an Age of AI, with extra shows on Copyright law and Blockchain, as well as audiobooks and voice technology also coming shortly. There are links in the show notes and clickable links in the ebook of you want to investigate further.
If the technology language sometimes goes over your head, think of this as an exercise in awareness. We can only learn new things if we learn new language to express a new reality.
Every major industry on our planet is about to be completely reimagined
“Every major industry on our planet is about to be completely reimagined. For entrepreneurs, for innovators, for leaders, for anyone sufficiently nimble and adventurous, the opportunities will be incredible… Welcome to an era of extraordinary.”
Peter H. Diamandis and Steven Kotler, The Future is Faster Than You Think: How Converging Technologies Are Transforming Business, Industries, and Our Lives
On a societal level, the impact of these converging technologies go far beyond authors and publishing, but as a writer and a reader, I’m deeply committed to the future of books, so that will be my focus here.
It’s time to change our business model. If we embrace this wave of converging technology, we can create abundance in our industry, enabling new forms of creativity, growing the market with new products and experiences, and expanding revenue for the entire supply chain.
We are creators. We turn ideas in our heads into books in the physical realm.
We can use these technologies to surf the wave of change and invent the decade ahead — together.
I hope you will join me on the journey.
Introduction
“People are usually afraid of change because they fear the unknown. But the single greatest constant of history is that everything changes.”
Yuval Noah Harari, Homo Deus: A History of Tomorrow
2016. Alpha Go, DeepMind’s AI, beat world champion Lee Sodol at the game of Go. One move was so shocking and original — and beautiful — that many saw this as the moment when AI truly moved into the creative realm.
2019. Go champion Lee Sodol retired from the game, saying that, “AI cannot be defeated.”
2020. A Chinese court granted copyright protection to written work generated by an AI writer.
Artificial Intelligence is already embedded in much of our daily lives and it’s increasingly moving into realms that impact authors and the publishing industry. Ignoring it or denying its potential power will not make it go away. We need to embrace the opportunities and engage in conversations around possible threats in order to reinvent our industry for a very different future.
The pandemic of 2020 has accelerated converging technologies and changed human behavior across the globe to favor digital business models.
In this book, I discuss current technological and societal trends and consider the opportunities for authors and the publishing industry over the next decade.
Why this book now?
I sit in the overlap between the creativity of authors, the business of publishing, and the technology of artificial intelligence. I remember when Alpha Go beat Lee Sodol. I watched his face as realization dawned and I began to devour books, podcasts, and articles on AI and its impact on my industry.
I’m an independent author and creator — I write thrillers as J.F.Penn, and non-fiction under Joanna Penn. I’m also a podcaster, business owner, tech enthusiast, and amateur futurist.
I spent thirteen years as a management consultant implementing IT systems, working for companies like Accenture and Cap Gemini in the UK, Europe, and Asia Pacific. Part of my job was Business Process Re-engineering (BPR). I would assess the current state business model and design future state processes to improve efficiency. I worked with programmers to implement those ideas, designing interfaces with external systems, and configuring software to meet business needs.
I’ve also run my own businesses for the last twenty years, and in the last decade, I’ve grown my author business on the back of emerging technologies. I’ve sold books in 159 countries and they are available in ebook, audiobook, paperback, hardback, and large print editions. My award-winning podcast, The Creative Penn, has been downloaded over five million times in 223 countries. My business is global, digital, mobile, and scalable by design.
I am usually early in adopting technologies that relate to authors and the publishing industry. I first self-published in 2008, back when it was considered ‘vanity’ instead of a savvy business decision by an empowered creative entrepreneur, and before the Amazon Kindle brought ebooks to the masses. I started my podcast in 2009, a decade before the big money entered podcasting, and publishing acknowledged its important role in book marketing.
As a writer and a reader, as someone deeply committed to the future of books and the creative industries, I want to help authors and the publishing industry surf the wave of change, rather than drown in it.
This is not the future. This is now.
The term ‘artificial intelligence’ has become almost as meaningless and all-encompassing as ‘the internet’ or ‘electricity.’ There is a form of artificial intelligence in much of what we do now, especially in this pandemic year, which has driven so many more online all over the world.
When you use Google to search, or Facebook to connect with friends, or shop online with Amazon, you use different forms of narrow artificial intelligence. When you watch Netflix, or YouTube, or listen to Spotify, you train AI algorithms with your behavior. Digital transformation is accelerating in almost every domain, and AI lies at the heart of this shift.
This book is an attempt to pull together multiple converging technologies and apply them to our industry, one that is known for its resistance to digital change. It is a positioning paper with the intention of exploring these issues further in more detail.
The initial chapter focuses on writing in the age of artificial intelligence, where AI can be a co-creator through Natural Language Generation tools. The next chapter goes into the impact of AI on copyright law, and the potential of Blockchain technology for licensing. These are the most significant aspects facing authors and publishing.
Subsequent chapters explore the potential of AI-assisted translation, voice technologies, virtual worlds, and augmented reality, as well as the societal and demographic shifts that will open up a truly global market.
I’ve linked to source articles as much as possible within the text and there are more books and resources in the Appendix, but of course, this is a vast field and I can barely scratch the surface in this short work. Chapter 2 on copyright law touches on bias and diversity, but I leave it to others to cover the ethical and moral issues of AI in detail.
I’m a techno-optimist, and after researching these areas, I’m more excited than ever to be an author entrepreneur. The opportunities for creators continue to expand — but only if we embrace the future, not run from it.
Writing in the Age of AI
The release of Open AI’s GPT-2 Natural Language Generation model in July 2019 prompted my original blog post and podcast episode on 9 Ways That Artificial Intelligence (AI) Will Disrupt Authors and Publishing in the Next 10 Years, which listed all the ways AI was already starting to impact the domain of writers.
Technology inevitably moves on and in May 2020, Open AI released GPT-3, over 100 times larger, trained on vastly more data and much more powerful in terms of what it could generate in terms of language-based output. There are now far too many examples of AI creativity to list individually, so this is merely a taster of the current state and my reflections on the possible impact. For a deeper look, I recommend The Artist in the Machine: The World of AI-Powered Creativity by Arthur I. Miller, and check out #creativeai on Twitter.
As a creative, it’s fascinating to consider the artistic and playful possibilities of AI collaboration, but as someone who makes a living from my words, the commercial impact is just as important. I’ve covered both in separate sections below.
What is GPT-3 and why is it such a big deal?
Generative Pre-trained Transformer 3 (GPT-3) is a language model that uses deep learning to generate human-like text. It ingests huge text-based datasets and outputs anything with a language structure based on prompts. It can write poetry or narrative text, answer questions, translate, and even generate code. There is no need for programming or human supervision. The AI trains itself based on the data ingested and generates original output.
As John Thornhill noted in the Financial Times in August 2020, “What makes GPT-3 astonishingly powerful is that it has been trained on about 45 terabytes of text data. For comparison, the entire English-language version of Wikipedia accounts for only 0.6 percent of its entire data set. Or, looked at another way, GPT-3 processes about 45 billion times the number of words a human perceives in their lifetime.”
For some examples of output, check out Gwern Branwen’s investigations, which include dialogue, humor, poetry, and even art criticism. They note that, “GPT-3’s samples are not just close to human level: they are creative, witty, deep, meta, and often beautiful. They demonstrate an ability to handle abstractions, like style parodies, I have not seen in GPT-2 [the previous model] at all. Chatting with GPT-3 feels uncannily like chatting with a human. I was impressed by the results reported in the GPT-3 paper, and after spending a week trying it out, I remain impressed.”
There are many valid criticisms of GPT-3, but in looking ahead, I think in terms of GPT-X, that unknown point where later versions of the model become far more sophisticated. As Daniel Susskind notes in A World Without Work, “Machines do not need to replicate human intelligence to be highly capable.”
These technologies are heading for the mainstream
Tools like GPT-3 and other advanced Natural Language Generation applications are still primarily in the research and development phase, but in September 2020, Open AI licensed GPT-3 technology to Microsoft. The API remains in limited beta at the time of writing, but we are likely to see commercial applications in the coming years given its potential in so many domains.
In the last iteration of technological change, businesses moved into the cloud with Software As A Service (SAAS). Think of the companies you use: Entertainment options like Netflix, Amazon Prime, and Audible; business and collaboration tools like Dropbox and Slack, and publishing platforms like Amazon KDP, which all function in the cloud. These span the digital and physical worlds. For example, I can upload a file to Amazon KDP Print or Ingram Spark and order a physical copy of my book, which is printed and delivered the next day.
The next iteration of technological change is AI-As-A-Service, where businesses will use options such as Amazon’s AWS, Google Cloud AI, or Microsoft Azure to build new offerings, transforming existing industries and inventing new ones. As the internet revolutionized business in the first two decades of the twenty-first century, these AI-based solutions will transform the next.
This is not years away. This is happening right now. Just go to the home page of these services to see what they already offer in terms of AI assistance: Natural Language Processing, Document analysis, and text extraction, Real-time translation, Text-to-speech, and much more.
Some say it’s too expensive, and machine learning, in particular, is resource-heavy, but there are technological developments that may solve these problems. Researchers have found a way to use light instead of electricity, which “can lead to substantial advancements in the speed and efficiency of the neural networks involved in machine learning. The approach also allows the AI to teach itself independently, without supervision.” [Big Think]
Quantum computing may also provide an entirely new framework. As WIRED magazine put it in March 2020, “Quantum computers aren’t just about doing things faster or more efficiently. They’ll let us do things that we couldn’t even have dreamed of without them. Things that even the best supercomputer just isn’t capable of. They have the potential to rapidly accelerate the development of artificial intelligence.” [WIRED]
So how should authors and the publishing industry approach such technological change?
Embrace curiosity and play with artistic collaboration
If you can put aside fear and doubt, it’s possible to see AI-augmented creativity as playful and revel in the possibilities. Artists in other mediums — musicians, visual artists, designers, poets, and more — are already co-creating with AI in a positive, experimental, and playful way to make new art.
Check out AIArtists.org for the world’s largest community of artists exploring the impact of AI on art and society. It includes musicians like Taryn Southern, who created the first solo album composed and produced with AI, and writer Ross Goodwin who co-created an award-winning novel with AI, as well as many others across diverse creative fields.
It’s exciting to consider the new artistic forms that might emerge. I’ve co-written several times and created new books in the world that I could not have imagined alone. How much more could we create in collaboration with a ‘mind’ that is not like our own?
Designer Es Devlin spoke at the WIRED Live conference (online) in November 2020 on the future of creativity. In Poem Portraits, she used 25 million words of nineteenth-century poetry to train a machine learning model, then merged it with words donated by a gallery audience of 1500 people. She called the result “a collaboration between the living and the dead.”
That sentence gives me chills, possibly because I write darker fiction and love the idea of collaborating with the dead! It also opens up ways for the writing community to collaborate far more than we do right now because new methods will emerge.
Es Devlin said that she “uses technology to create magic,” and this attitude to AI-augmented creation will drive innovation and add to the beauty of the artistic world. An exciting prospect for any creator!
Impact on the commercial business model for writers, authors, and publishing
As much as I love the playfulness of creative possibility, I currently run a multi-six-figure business based on the words I write and speak. My income depends on readers and listeners paying for and supporting my work. I can create all the art I want in the world, but at some point, I need to make money!
Journalists and freelance writers are paid for the words they craft into a narrative. Authors are paid for creating and licensing intellectual property assets (books). The business of publishing makes money from selling books to readers in various formats. We all have an interest in how AI impacts the commercial side of writing.
Journalism has been using text-generation AI tools for several years. Companies like Automated Insights and Narrative Science turn data into news reports and business content for media and corporations. [Forbes] In February 2020, Reuters took this even further by developing a “fully automated, yet presenter-led sports news summary system” using an avatar of a presenter “reading a script generated from real-time match data.” [Forbes] You can even make your own at Synthesia.io.
Automated Insights “generates human-sounding narratives from data, making it easy to produce real-time, written analytics, personalized reports, and stories at scale.” Their natural language generation platform, Wordsmith, is used across 50 industries and their website says it can “generate millions of narratives in a matter of milliseconds with the Wordsmith API.”
Janelle Shane explains how this works in You Look Like a Thing and I Love You: How Artificial Intelligence Works and Why It’s Making the World a Weirder Place.
“AI-generated writing allows news outlets to produce the types of articles that were formerly cost-prohibitive. It requires a human’s touch to decide which articles to automate and to build the AI’s basic templates and stock phrases, but once a paper has set up one of these hyperspecialized algorithms, it can churn out as many news articles as there are spreadsheets to draw from… Human reporters can spend their valuable time on creative investigative work instead.”
But what about the articles that so many freelance writers create as a primary source of income?
In August 2020, college student Liam Porr used GPT-3 to write an article: “Feeling unproductive? Maybe you should stop overthinking.” He said, “it’s possible it could pass as content marketing.” It reached the number one spot on Hacker News before he revealed its provenance. [The Verge]
In September 2020, The Guardian worked with Porr to input prompts into GPT-3 and generate material for an article. They published the process of co-creation in How to edit writing by a robot: a step-by-step guide.
They concluded, “The Guardian could have just run one of the essays in its entirety. However, we chose instead to pick the best parts of each, in order to capture the different styles and registers of the AI. Editing GPT-3’s op-ed was no different to editing a human op-ed. We cut lines and paragraphs, and rearranged the order of them in some places. Overall, it took less time to edit than many human op-eds.”
This is the likely way we will work together with AI Natural Language Generation tools. As in centaur chess, where a human works together with an AI, we will become centaur writers, working together to create something more powerful than each could produce alone.
As Kevin Kelly says in The Inevitable: Understanding the 12 Technological Forces That Will Shape Our Future,
“This is not a race against the machines. If we race against them, we lose. This is a race with the machines. You’ll be paid in the future based on how well you work with robots… It is inevitable. Let the robots take our jobs, and let them help us dream up new work that matters.”
What about books?
The first AI-generated textbook was published in April 2019. The Verge reported, “Writing in the introduction, Springer Nature’s Henning Schoenenberger (a human) says books like this have the potential to start “a new era in scientific publishing” by automating drudgery.”
But while the algorithms can generate chunks of text, they have an issue with longer narrative. As Janelle Shane says in You Look Like a Thing and I Love You, “As of 2019, only some AIs are starting to be able to keep track of long-term information in a story — and even then, they’ll tend to lose track of some bits of crucial information. Many text-generating AIs can only keep track of a few words at a time.”
But that was 2019.
Part of the problem was that the language models could only ingest a few paragraphs of text at a time, but in January 2020, Google reported that their AI language model Reformer could ingest up to one million words, easily devouring Dostoevsky’s Crime and Punishment at 211,591 words. [Venture Beat]
Story structure is another critical aspect. While AI poetry can be beautiful, it is short form and can be abstract by design. A longer, coherent narrative arc is another question entirely. Which is why Scriptbook.io is so interesting because many novelists use screenwriting as a way to hone their craft, and Scriptbook “provides artificially intelligent script analysis, AI driven content validation and automated story generation.”
Their tool, Deepstory.ai, is an automated script generator using “neural networks to help creators co-write and co-create original stories with the help of our AI.”
Nadira Azermai, CEO of ScriptBook, said in The Guardian in January 2020, “Within five years we’ll have scripts written by AI that you would think are better than human writing.”
There are already novels written entirely by AI or in human collaboration with it. The award-winning novel, 1 The Road by creative technologist Ross Goodwin, was published in 2017. Publisher JBE Books describes the method of creation.
“As a prominent Artificial Intelligence creator, [Goodwin] outfitted a Cadillac car with a surveillance camera, a GPS unit, a microphone and a clock, all connected to a portable AI writing machine that fed from these input data in real time. Together, they traveled from New York to New Orleans, in an experimental automation of the American literary road trip. As they drove, a manuscript emerged line by line from the machine’s printer on long scrolls of receipt paper that filled the car’s rear seats over the course of their journey.
1 the Road offers the first real book written by an AI, which captures us from the first page, when the journey begins. Rooted in the traditions of American literature, gonzo journalism, and the latest research in artificial neural networks, 1 the Road imposes a new reflexion on the place and authority of the author in a new era of machines.”
BooksByAI.com is a bookshop for printed paperback books entirely generated by Artificial Intelligence.
“Through training, the artificial intelligence has been exposed to a large number of science fiction books and has learned to generate new ones that mimic the language, style and visual appearance of the books it has read. None of the stories, titles, descriptions, book covers or reviews related to any of the books on Booksby.ai has been written or designed by humans.” The books are for sale on Amazon, and some even have good reviews.
These books acknowledge their AI generation and have been curated in some way by humans who are, to some extent, putting their reputations on the line by associating with the AI label. The books are more artistic experimentation than commercial proposition. But what about those who might use this technology with no such restrictions?
In The Artist in the Machine, Arthur I. Miller writes about Philip M. Parker, “whose patent software can sweep the web, pick up information on virtually any subject, and turn it into books to be printed on demand… Parker has produced over two hundred thousand books. Some 90,000 are on Amazon… Parker claims he is now tailoring his algorithms to produce romance novels.”
The quality might be dubious now, but with developments in technology, human co-creation, and curation, that will change. It will increase the tsunami of available books many times over — and it is already hard to navigate.
So, let’s assume an AI can write a book, or at least output one from a Natural Language Generation model. If it satisfies the reader, will they care if it’s primarily or even entirely written by AI?
I don’t think so.
How many books have you read where you don’t know or care who the author is? Given that so many already write under pseudonyms, how do you know you are not already reading books written by algorithms?
If we accept that this will happen in the next decade, how will it impact authors?
Will AI tools make authors obsolete?
Of course not.
Most authors don’t write primarily for money. We write because we want to share our stories, to change people’s lives with our ideas, to entertain, to inspire, to educate. We write because we are creative and writing is how we create. We would do it even if there was no money in it at all.
But there is money in it.
This potential tsunami of AI-generated content is a concern. Independent authors understand the existing problem of scammers and plagiarism that plague the Kindle store on Amazon in particular, because of the possible financial rewards. There are already humans taking the hard work of authors and altering it to create plagiarised versions, as author-activist David Gaughran has covered on his blog. How much more will this happen when scammers can use machine learning and Natural Language Generation models?
But it’s not just scammers who have the potential to create books by AI.
The publishing industry owns and controls the largest dataset of all.
Books written by human authors, edited and curated by professionals, licensed and paid for, all separated neatly into genres.
As British mathematician and data scientist, Clive Humby, said, “Data is the new oil,” the wealth generator of the digital age — and publishers have a lot of it.
Here’s a thought experiment. A publisher of romance novels, the highest-selling of the genre fiction categories, has paid writers for many years to create books to a certain template. They have tens of thousands of those books, and they’ve also bought the backlist of bestselling writers. They use all that to create a dataset of stories that satisfy a particular target market. The dataset is ingested into something like GPT-X and new romance novels are generated which they edit and publish under their imprint.
If I was running a publishing company and sitting on that much intellectual property, I’d be looking at this. I’m not a copyright lawyer, but from what I have researched so far, there is nothing legally preventing this from happening right now.
Most authors don’t truly understand intellectual property rights licensing and therefore under-value the worth of their books over the long-term. I know authors who have licensed their work for “All languages, all territories, all formats existing, and to be invented, for the term of copyright” (up to 70 years after the death of the author) for a few thousand pounds. In considering the thought experiment above, even this model could be under threat.
As artists, we will continue to create whatever the future of AI. But we need copyright reform and a new way to track and reward intellectual property in order to be paid and recognized for our creative work, which I will cover in detail in the next episode.
Conclusion
After much reading, listening, and research, this book is an attempt to corral my ideas around converging technologies into one place.
I know I won’t be right about it all, but everything I’ve written about here is possible and there are companies in almost every area working on solutions. There is much to do to turn our creative, technological future into reality.
My mission for over a decade has been to empower authors with the knowledge they need to choose their creative future, and I intend to continue that with this new direction, sharing my experience and lessons along the way.
I hope you’re inspired by the possibilities of the next decade. Please do leave your comments or questions here on the episode. If you work in these areas and would like to connect, or if you have new ideas that might empower authors and propel our industry forward, please reach out.
I’ll also be speaking professionally and consulting in this area, so if you have business opportunities, I’d love to talk. Contact me here.
You can now get the ebook on all the usual stores and that covers a lot more ground. Coming soon in print and audiobook. I’ll be back with another solo show covering copyright law, blockchain for smart contracts, and micro-payments. Happy writing, and I’ll see you next time.
The post Writing In An Age Of Artificial Intelligence (AI) first appeared on The Creative Penn.
Go to Source
Author: Joanna Penn