“AI is not another incremental shift,” writes Javier Celaya. “It touches every part of what publishers do—creation, curation, production, translation, marketing, distribution, and rights management.”

By Javier Celaya

There is a growing level of AI literacy among publishing professionals. But while many are daily users of chatbots and some are experimenting with AI-assisted copywriting, proofreading, or translation tools, very few of them have created an AI agent to redefine an existing business process, or used a LLM to evaluate and compare different business scenarios. Knowing how to use an AI tool is not the same as understanding it—and that gap, I have come to believe, is one of the most consequential blind spots in our industry today.

After two decades working in publishing and creative industries, that conviction is what led me to enroll this year in an executive master’s program in artificial intelligence at the Instituto de Inteligencia Artificial (IIA) in Spain. It was not a decision I took lightly. I spent several weeks comparing programs, reading curricula, and speaking with alumni before committing. What I was looking for was not a crash course in prompting, or an overview of AI tools, but a rigorous understanding in how these systems actually work—how models are built, how they are trained and retrained, what structured data really means for tech companies, who owns the data, and where the hard limits of current AI models genuinely lie.

Learning the Language Underneath the Language

The program changed the way I think, talk about, and use AI—completely. Previously, I was, at best, an entry-level user of several tools. Now before opening any AI tool, before writing a prompt, before touching a tool, now I think about exactly what I want the AI agent to do for me, not “in general,” but specifically.

I also now find myself approaching publishing-industry AI conversations with a fundamentally different frame. When colleagues speak about the risks of AI-generated content, I understand the probabilistic mechanics behind why those risks exist. When vendors promise that their model is “trained on clean data” or try to avoid talking about “ownership of data sovereignty or data expiration dates” or “reuse policies” or “data access levels”, I know which questions to ask. This kind of informed skepticism is worth more than mastering any single tool.

I found particularly relevant the second half of the program, which examined AI adoption across sectors: healthcare, automotive, pharmaceutical, and entertainment industries. Studying how other industries have navigated the integration of AI into their core processes, where they stumbled, which use cases delivered real value, what governance structures emerged, gave me a comparative lens that the publishing industry needs take into consideration to move forward. In publishing, we tend to talk about AI from a defensive position, while the rest of the world has already kicked off projects using innovative tools that we are still debating.

Beyond Cost Savings: AI as a Growth Engine

The publishing industry’s focus on AI is almost entirely stuck in saving costs. Most publishing entities are looking to deploy AI to trim costs, but these processes represent the floor of what AI can do for publishing, not the ceiling.

The strategic questions that will define the next decade are barely being asked: Which processes in a publishing house can be fully automated? Which require a human at the beginning, or the end, of the workflow? Which must remain entirely human? Most critically: how can AI help publishing grow, rather than simply shrink its cost base?

For example, AI-powered translation and production of ebooks and audiobooks in multiple languages could bring key titles and authors to entirely new audiences worldwide, at a scale and speed previously impossible. The redefinition of marketing and distribution—from local to genuinely global—becomes conceivable when language, storytelling contextualization and marketing localization are no longer the bottleneck. The opportunities are substantial, yet few entities in our industry are analyzing these possibilities with the deepness they deserve.

Partners, Adversaries, or Both?

Meanwhile, the most difficult question in publishing right now is what relationship the industry should have with AI companies: partners? or adversaries?  My answer, after serious reflection during the master program, is both.

There is no intellectually honest way to avoid the fact that many AI companies built the capabilities that now threaten and attract publishers alike by training on copyrighted works without consent or compensation. The legal and moral case for seeking redress is strong, and publishers should pursue it. Lawsuits and business partnership negotiations are not contradictions—they are parallel tracks that can and should run simultaneously.

Publishers and AI companies should define together the collaboration scenarios, reimagine new distribution channels, as well as the audience relationships that will define how content is discovered and consumed in the coming years.

Education Is a Strategic Imperative

I would go further than recommending that publishing professionals continue their AI education. I would argue it should be a requirement for anyone in a managing position. No managing director can responsibly define an AI strategy for their publishing house or streaming platform without understanding, at a structural level, what AI can and cannot do.

The digital transformation of publishing has been underway for two decades, and the sector has repeatedly underestimated the pace at which new technologies redraw the competitive landscape. AI is not another incremental shift. It touches every part of what publishers do—creation, curation, production, translation, marketing, distribution, and rights management.

The leaders who will navigate the AI age most successfully will not be those who understand AI in an abstract sense, but those who understand it deeply enough to ask the right questions— of their teams, of their technology partners, and of the models themselves.

Javier Celaya is founder of Dosdoce.com, a company launched in March 2004 to help the cultural & entertainment sectors understand the digital age we are living in. Throughout the years, Dosdoce.com has compiled more than 300 studies and reports on the use of new technologies in different areas of the cultural sector (content creation, business models, distribution, marketing, etc.).