AI Strategies for Book Publishing Companies

Here’s a draft roadmap for implementing AI in a book publishing company.

1. AI Adoption Decision: Decision-makers need to understand the fundamentals of AI as it impacts publishing, and to be willing to make a series of decisions about AI within their organization. If you decide to move forward with AI, there’s no point in moving tepidly. You’ll learn little from gingerly poking AI around the edges — you’ll get the impression that it offers little value. The only useful approach is to go all in.
2. Digital Readiness Audit: Before you move forward with specific AI initiatives, you need to get your “digital house” in order. Any existing workflow problems that your publishing company still faces should be addressed NOW, because the company needs to approach AI from a position of digital operational strength. Is your metadata complete, accurate and up to date? Are all of your backlist books digitized, with the content consistently tagged? Are you efficiently using a sufficiently robust TMS and/or CMS (title management system or content management system)?
3. Staffing for AI: See section below.
4. AI Training: See the section “Some Additional Sources” below for training resources beyond this book.
5. Pilot Project: Plan for a pilot project to be implemented by the early enthusiasts. Don’t make it too easy, but make sure it’s a win. Create a framework for evaluating the project; determine success metrics (e.g., improved efficiency, cost reduction, or enhanced user experience).
6. Financial Planning: You will need to budget for AI training and implementation. Software costs are modest; nearly all of the costs are in training and recruitment. Larger initiatives will require the skills of consultants and outside vendors. Where will the funding come from for your AI initiatives? Can you project a ROI? I’ve created a somewhat arbitrary goal for book publishing companies looking to adopt AI technologies: look to drive 15% out of your fixed costs while building the sales of backlist titles by 15%.
7. AI & Publishing Value Chain: Analyze how AI could disrupt various aspects of the publishing value chain, such as author discovery, editorial processes, book production, marketing, and distribution. How will these disruptions impact your organization? What actions do you need to take to minimize the effects of the disruption?
8. 1-2 Year AI Strategy: Create an internal task force to build out a 1-year and 2-year AI strategy for your organization. Anything past 2 years is wild speculation — arguably even with a 1-year and 2-year AI strategy. But in times of uncertainty, planning is an essential discipline. With each perceived opportunity, decide whether you are creating organizational efficiencies with AI (cost savings), or driving new revenue. Regardless of the anticipated dollars that might be saved, or the revenue that might be gained, assume that both types of initiatives have equal value. Nonetheless, set targets and timelines. Given the speed of change for AI, include quarterly review/adjustment cycles.
9. AI Quality Control: Establish metrics for evaluating AI output quality across different departments. Define acceptable error rates and correction procedures. Create workflows for human review and oversight.
10. Industry Allies: Figure out who your allies are, whether at other publishing companies or at trade associations, or just enthusiastic individuals/analysts/consultants. Reach out and confer. Hire outside talent to analyze, to advise, and to assist. Attend industry conferences and online webinars.
11. Vendor Assessment: Make a first assessment on which vendors or other service partners might be able to provide software or systems-based approaches to move the organization forward with AI technology.
12. Competitor AI Analysis: Assess what your competitors are doing with AI. What outcomes do you expect them to achieve from their efforts? How should this impact your strategy?
13. AI Policy Development: See next section for more.
14. Data security and privacy: Implement safeguards for the use of AI tools for manuscripts and associated data. Ensure GDPR and CCPA compliance when using AI. Evaluate other regulations within the U.S. federal and state governments.
15. AI Risk Management: Create contingency plans for AI program failures. Develop protocols for AI-related PR issues. Plan for potential AI regulatory changes.
16. Author Relations: Create support systems for authors using AI tools. Develop AI collaboration guidelines. Plan a communication strategy around AI initiatives. Consider programs to educate your authors on the pluses and minuses of engaging with AI.
17. AI Use Detection, Fact-checking and Plagiarism Detection: Invest in understanding which AI tools can detect AI use, plagiarism or inappropriate content. Explore AI use for fact-checking.
18. Accessibility and AI: AI tools are proving to be fast and largely accurate for creating accessible content, including both alt-text and audiobooks. How can you take advantage of these tools?
19. AI for Marketing: AI tools for marketing have tremendous potential and are being widely used across multiple industries. Your staff need to prioritize AI adoption across marketing and sales functions.
20. AI & Audiobooks: The technology for synthetic voices works today, and is essentially undetectable. Determine your strategy on the use of AI for audiobooks, particularly for backlist.
21. AI & Translation: AI translations using large language models greatly increase translation efficiency. What does this mean to your strategy of selling and buying foreign-language rights? Which of your books could you translate to Spanish, just for the U.S. market?
22. The Long-Term Vision: While acknowledging the limitations of long-term predictions, brainstorm some visionary thinking about the future of AI in publishing, including potential scenarios and their implications for the industry. Encourage an internal team of enthusiasts to explore speculative AI innovations, including AGI (artificial general intelligence).

Despite its widespread use, few publishers have publicly defined their AI policies, and communicated their approach to AI to the public. The term ‘the public’ has a slippery significance here, when you consider the different publics addressed by trade, scholarly and educational publishers.

For trade publishers the most important audience is authors and their agents. Scholarly publishers face different obstacles, when they consider AI’s promising impact on research, and then AI’s more problematic impact upon converting research into narrative (Avi Staiman wrote a thoughtful post on this topic). For educational publishers, establishing policies is tricky, as AI’s encroachment on the practice of teaching, of education, is multifaceted and complex.

Publishers face two big challenges as they move forward with AI technologies. The first is to develop a corporate position about how to approach AI generally, on how to incorporate AI into their workflows. The second challenge is communicating their position, clearly and unambiguously, to their constituents.

The publisher policies I have seen are mostly flawed. Some of them are in fact policies directed externally, at authors, with a range of admonitions about what is acceptable practice (not much) and what is not acceptable (lots). O’Reilly’s “AI Use Policy for Talent Developing Content for O’Reilly” goes on for pages and pages, with esoteric guidance, such as “DO NOT use any OSS GenAI Models that produce software Output that is subject to the terms of a copyleft or network viral open source license.”

On the other hand scholarly publisher Elsevier, in the “Elsevier Policies” section of its website, includes statements on “Responsible AI Principles,” “Text and Data Mining,” and “The use of generative AI and AI-assisted technologies in writing for Elsevier.”

The few internal, unpublished, publisher policies that I’ve seen are conservative, excessively so. These publishers reacted too quickly to the range of perceived and possible threats, and to their authors’ anxieties, and have hamstrung their own ability to engage robustly with this fast-developing, fast-changing technology.

It’s a given that they will use AI ‘responsibly,’ whatever that means. It’s a given that they have the utmost concern for authors’ intellectual property and for aggressively protecting author’s copyrighted work. (Although, of course, these principles must be declared publicly, and often reiterated.)

But what else?

• Will they allow AI to have a role in editorial acquisitions? Can AI take a look at the slush pile?

• Will they allow AI to have a role in developmental editing, line editing and copyediting?

• Will they allow AI to have a role in determining print runs and allocations?

• In creating accessible ebook files, including alt-text?

• In aiding audiobook creation in cases where it’s not economically-realistic to hire talented human narrators?

• In aiding foreign language translation into markets where rights would never be sold?

• In developing marketing material at scale?

• In communicating with resellers?

If so they must make this clear, and clearly explain, the thinking behind these policies. Publishers must be brave in countering the many objections of most authors at this time of fear and doubt.

Management can’t tackle AI on their own, and they can’t tackle it just with outside vendors and consultants. Your success in executing your overall AI strategy will be determined by whether you can get buy-in from your knowledgeable and experienced publishing team.

Some of your staff have already made progress on their own, probably without your support or explicit endorsement. Try to identify the enthusiasts on your staff. …strategic direction?

At the February 2024 PubWest conference in Arizona a speaker from outside the publishing industry suggested that one of the uses for AI will be replacing interns. The room burst into flames. She meant well—indeed an April 10, 2024 report in the New York Times describes how Wall Street investment banks are looking to replace many of their interns with AI. Similar to the case in publishing, an obvious concern is: how do you find senior analysts if they can’t start off as junior analysts?

The publishing industry has always relied on internships. A 2019 study found that 80 percent of the people who had worked in publishing for less than fifteen years had previously interned.

In part it’s a way to get the grunt work dispatched at a reasonable cost. But that pales against the larger reality that no publishing school can equip someone to join a publishing company at the level of middle-manager. The only way to develop the skilled staff of tomorrow is to train interns and apprentices today.

The objective here is not to seek to replace interns with AI, but instead to make their work more productive and rewarding using AI tools, benefitting both the intern and the publishing company.