A Look at AI in Publishing. Open The Pod Bay Doors, Hal.

Stephen Hawking warned the world in 2014 that Artificial Intelligence (AI) could be the end of mankind, such was the great man’s deep concern about machines surpassing humans. Movies like “2001: A Space Odyssey” (1968) – in which the iconic phrase “Open the pod bay doors, Hal” was first delivered – and Blade Runner (1982) had opened a pang of anxiety in our minds. So, when IBM’s Deep Blue computer defeated Garry Kasparov, the best chess player in the world in 1996, we were reminded that we should be worried.  

The term “Artificial Intelligence” has been around for quite some time, having been first coined in 1956 at Stanford University by computer scientist, John McCarty. In late 2022, however, we got the latest AI jolt from the amazing antics of the internet chatbot – ChatGPT. The app from OpenAI in San Francisco allows users to create (among other things) high-quality text from only basic prompts. It is currently free, so that OpenAI can gather feedback and recommendations from the audience. And judging by the frustrating backup of people trying to get into the app every day, this is a large and enthusiastic audience. 

On the face of it, this appears to be a game changer. If you are used to reading articles about AI, you will be aware that they have all hitherto assured us that AI will never replace human writers and the creative writing process. Well, it appears that “never” is here.  

How does AI impact the typical medium-sized publisher?

Human-To-Machine (H2M) communications engineers talk about an “H2M” scale of 0 to 4, where level 0 is all-human activity and level 4 is all-Machine. Levels 1 through 3 are intermediate steps, and most current AI solutions are at Level 1, so we are clearly at the beginning of an exciting journey.  

In this article, we take a look at whether AI is going to change any lives or put us all out of work. Being focused on publishing, we investigate what impact AI has on consumers and on the typical medium-sized publisher, and how they can put all this smart technology to work.  

As consumers, we all see AI in use every day although we may not even recognize it as such. Any time we see a “recommendation” from a company such as Amazon, or Netflix or Spotify, we are seeing AI in action, based on the data they have gathered on our own preferences and behaviors. Simple AI apps are commonplace. To give the reader an idea of how commonplace, see the end of this article, for a sample list of publishing apps from a brief internet search. Many have the top-level domain name “.ai.” 

Human vs. Computer generated writing 

Even though ChatGPT and tools like it offer some striking prose, most of us would still say that the difference with human writing is that humans can put together unmatchable written structures. What makes our creations valuable is how humans connect nouns and adjectives and paragraphs in new and unusual combinations. Each combination is totally new. We like to say we have “feelings” whereas machines are just logical and rational – they can only use what they have been fed (both good and bad). You would think that to reproduce all the vagaries of skilled writers’ minds, the machine training system must learn so much material that it is beyond even the most capable computer systems that we have today. Well, we are not there yet but it is likely that we will get there. It is only a question of time. 

Uses of AI in Publishing Today 

Some of the early publishing applications automatically edited (and translated) incoming manuscripts. Today, the use of AI has spread to most of the functions of the commercial publisher: in marketing, customer service, finance, editorial and production processes.  

Marketing is now possibly the biggest single user of AI in publishing. They often use AI to analyze reader preferences and behaviors for audience building, resulting in automated personalized marketing (via email and social media). It saves a lot of time and is proven to be much more effective than a one-size-fits-all approach to building community with readers. Text auto-tagging and SEO (Search Engine Optimization) techniques can easily create the right hashtags to enhance the probability of discovery by consumers and advertisers. They are the building blocks of a good digital marketing strategy. 

In Customer Service Departments, a great deal of customer goodwill is lost, waiting to speak to a human customer service representative, and so a well-prepared chatbot is reliable, factual, inexhaustible, and when the going gets rough, it can easily transfer control to a human, who will undoubtedly be less stressed because of the bot’s intervention. Reader retention is at stake. 

Finance departments are beneficiaries of AI applications for royalty contract analysis and creation, in generation of public reports, and for copyright and plagiarism checking. IP (Intellectual Property) is the publisher’s lifeblood. 

Editorial functions use AI in automated text formatting and analysis at manuscript acquisition, in copyediting, and proof reading, and even in the creation of attractive titles and cover art. AI data tells them which art and titles worked and which did not. In production planning, publishers can identify trends from reading data to modify content or to reorder their publication plan. AI also plays a big part in format conversion (from print to e-book or audio, for example), and in metadata management for ONIX input and export. 

Real Business Value – But how to get there

In most cases, the time saved in reducing manual effort in performing repetitive, low level but essential chores can be transformative for publishers. By moving the humdrum work to the computer, the result is a higher quality process. It allows humans to focus on less repetitive, more insightful tasks where humans excel, and leaving the large volume of simple repetitive tasks to the computer, in which it also excels. Thus, the biggest benefits of AI and machine learning to any publisher are the significant efficiencies in workflows and major reductions in time to edit, produce, and get the book to market in a timely fashion. There is genuine business value there. The benefits are unquestioned and often result in performing work at what might have been previously unimaginable rates. The question is how to implement AI successfully?  

It is true that any AI app is only as good as the training data that it has been fed. That, and the continuous updating of the data. It is likely that the typical medium-sized publisher cannot hope to collect the data economically or recruit the resources to do this themselves. For this reason, many AI applications are acquired rather than built in-house, and are best approached with an experienced external practitioner or consulting organization to provide the needed guidance, especially for the first project. 

It is important to know where and how to start. There is no question that a successful AI project is dependent on having a clearly defined objective, and careful identification of the problem that needs to be resolved. To that end, the project must have a well-defined scope, with metrics to define success (or otherwise). It should be something that has real value and ROI to the whole company and to which the important contributors are committed. And it is essential to start with a simple, independent application that can result in a quick win, not just to build excitement for later AI projects, but also to gain corporate experience before taking on larger, more complex implementations.  

For the future, the decisive question is where the content lies and how it can be made available. The book, as we know it, is here to stay. Everyone seems to agree that there will be publishers and editors in the future who will ensure the quality of published content on curated channels, but your guess is as good as ours as to who will be writing it.  

If you want to remove the guesswork and start a specific AI project but do not know where and how to start, feel free to get in touch with me at smayeres@knk.com, and I will be glad to help direct you to the right resources.


Examples of Everyday AI Apps:

    • krisp.ai – Noise cancelling app for removing background noise during calls
    • beatoven.ai – AI music generator for creating original tracks
    • cleanvoice.ai – AI-based voice enhancement tool for improving audio quality
    • podcastle.ai – Podcast hosting platform for creators
    • flair.ai – AI-powered customer service platform
    • illustroke.com – Tool for creating custom illustrations using AI
    • patterned.ai – AI-powered design tool for creating patterns
    • stockimg.ai – AI-based stock image search engine
    • copy.ai – AI-powered copywriting tool
    • copymonkey.ai – AI-based copywriting platform for social media
    • ocoya.com – Cloud-based project management tool for remote teams
    • unbounce.com – Landing page builder for businesses
    • vidyo.ai – Video analytics platform for businesses
    • tryMaverick.com – Business intelligence tool for sales teams
    • quickchat.ai – AI chatbot platform for customer service
    • puzzlelabs.ai – AI-based platform for data analysis and visualization
    • soundraw.io – AI music creation platform for non-musicians
    • cleanup.pictures – AI-based photo editing tool for enhancing images
    • looka.com – AI-powered logo maker for businesses
    • TheresAnAIforThat.com – AI-based platform for improving customer service
    • synthesia.io – AI video creation tool for businesses
    • otter.ai – AI-based transcription tool for recording and note-taking