richardobeirne.com

The annual conference for Irish publishers (AIPC) is a one-day event, run very professionally by Publishing Ireland, which took place this year in the Hyatt Centric hotel in The Liberties area of Dublin. Over 100 attendees included a group of students from University of Galway’s MA in Publishing course and a good number of visitors and speakers from the UK.

The advantage of a small conference is that there is only one track. Everyone attends every presentation. Some might not naturally choose a presentation on ONIX or EU sustainable paper sourcing regulations or the intricacies of post-Brexit distribution. It’s a bit like a crash course in publishing in Ireland. By the end of the day everyone leaves with a comprehensive understanding of the current state of Irish publishing – which seems to be generally in good shape.

I generally work in academic and research publishing rather than trade publishing which was the focus of this event. But it’s sometimes useful to come at familiar questions in a different context to see connections. A couple of thought-povoking conversations from the event have stuck in my mind…

• Academic publishers are currently grappling with research integrity – how to ensure that the content of research publications is valid and trustworthy. There seems to be a rapidly-growing question of publishing integrity in trade publishing, given that Generative AI combined with highly efficient digitized systems for ebook, print on demand, or audio book creation have reduced the barriers and timelines for getting a book from concept to distribution. Both industries are playing catch-up with technology. Only a third of publishers have AI policies for authors. Amazon has revised its guidelines for Kindle Direct Publishing to require (voluntary) declaration of AI content and distinguishes between AI-generated and AI-assisted content. It even aspires to “not allow content that’s typically disappointing to customers”.
• Cross-industry metadata initiatives have transformed print and digital publishing in the last couple of decades. Thinking about research integrity in academic publishing, is there a missing piece around metadata to validate research integrity in academic publishing?

Below are notes from the main #AIPC25 presentations:

• The Future of Publishing in a Multi-Platform World: Abi Watson, Enders Analysis
  • “Everyone (else) is miserable”. According to Enders Analysis data, newspaper circulation, broadcast television, even gaming and big tech revenues are falling or flat.
  • All content suppliers are facing the same issue of reintermediation and disintermediation by the GAMA platforms.
  • Licensing for book to screen adaptations are increasing, particularly on Netflix. Spotify is actively reducing its dependence on major music labels through a multi-pronged strategy that includes diversifying into non-music content like podcasts and audiobooks. [It was very noticeable that Spotify had one of the largest stands at the 2025 London Book Fair].
• Legal Issues for Publishers Update 2025: Samantha Holman, Irish Copyright Licensing Agency
  • Copyright has been called the Original Sin of AI.
  • EU copyright regulations review in 2026 is due to include a W3C machine-readable mechanism to streamline the communication of TDM rights and licenses.
  • The European Accessibility Act is “good for new content, challenging for older content.”
  • The Simon Mellins podcast was recommended for its coverage of accessibility in publishing, and more.
  • Accessibility has become a reputational risk for publishers with instances of non-compliance receiving public attention or legal action.
  • ICLA’s summary of the current market is that STM is “advanced” in compliance; Education publishing is “high legal risk and behind”; Trade publishing is “lower legal risk and behind.”
  • A “disproportionate burden” clause may offer legal cover for non-compliance in deep backlist content, but it’s important to record any reasoning in terms of accessibility actions deliberately not taken in case of future public or legal inquiries.
  • The ICLA is carrying out a consultation on collective GenAI licensing for its rights community.
  • In research publishing, the SCOIR project is working to draft secondary rights legislation for Ireland, creating an Open Access and Rights Retention policy framework for institutions and funders in Ireland.
• Three Years of AI in Publishing: George Walkley (Outside Context)
  • Generative AI gives rise to paradoxes: a technology that exceeds human performance against a variety of benchmarks, but still makes elementary errors and generates hallucinations; increasingly integrated into software, but doesn’t behave like traditional software.
  • Publishing – an industry that has successfully ridden previous technology waves – needs to make sense of the many AI paradoxes. However, a significant challenge is AI’s “capability overhang” (the gap between what AI technology is theoretically capable of achieving and what organizations can actually implement). As with all technology shifts, deriving value from AI requires people, culture, and processes as much as technology.
  • Two thirds of publishers don’t have a formal AI policy. Some large publishers are communicating AI (and its uncertainties) well with authors: AI guidance from Wiley and David & Charles were cited. However, there is need for realistic expectations in terms of what commitments publishers can commit to in terms of author contracts.
  • The issues of AI in publishing are well-rehearsed, but publishers’ focus is now on identifying valuable use cases for AI implementation. Pearson were cited as a large company implementing AI across a wide range of internal and product use cases.
  • In the AI world, you’ll learn more from experimentation than documentation.
  • For AI in image creation, Adobe Firefly is notable as commercially safe from a copyright perspective: It is trained on licensed Adobe Stock images, public domain content, and content licensed to Adobe.
• Metadata: a beginner’s guide. Graham Bell, EDITEUR
  • Standards organizations are in the “chaos minimization” business. Standards solve business problems and so must evolve alongside the businesses they support.
  • ONIX is the global standard language of the metadata supply chain.
  • Research has shown that the more information that customers have about a book, the more likely they are to buy it.
• EUDR: European Union Deforestation Regulation. Lisa Faratro, CPI Group
  • The requirement for wood and paper based products to provide a machine-readable ‘due diligence statement’ regarding sustainability. Publishing industry actors across publishers, printers, data, and metadata organizations have done huge work to be ready for compliance, but the regulatory rollout seems chaotic.
• Reviving the Backlist. New approaches to increasing availability, discoverability and sales of backlist titles. [Panel]
  • Digitization of workflows particularly advances in single copy, high quality, print-on-demand from PDF have given rise to ‘zero inventory publishers’ who effectively only hold copyrights and PDFs. The discovery, ecommerce, and distribution/delivery are often carried out by full service printing companies.
  • There’s an important difference in books that are out of print versus what could be called ‘out of commerce.’ When a book is declared out of print the copyright (usually) reverts to the author. With ‘out of commerce’ titles, the book exists in limbo – not available for sale but not reverting in copyright to the author to do anything alternative with.
  • In a globalized world, publishers need to be able to respond rapidly to demand. An example was given of a broadcast interview with an Irish author in the US. This created immediate demand for the title via Amazon which couldn’t be fulfilled. A conventional reprint of the title was arranged but was it too late?

Keep an eye on the Publishing Ireland website for details of next year’s conference – well worth attending!