AI and the Book Value Chain: Do We Need a New Concept of the Core of (Quality) Publishing?

Abstract

AI systems can now be helpful in pretty much every stage of the publishing value chain, the application use cases range from writing blurbs, the use for predictive pricing all the way “back” to the actual (co-)authoring of texts. This has analogies to the situation in the 1990s insofar, as early steps of digitization led to the irritating insight that elements that had been considered to be integral part of the essence of publishing (then e.g. printing on paper) were downgraded to contingencies. Thinking about this it becomes clear that the question what are the constitutive elements of publishing can already be asked concerning much earlier stages of book history, as will be shown with the help of a short introductory example. If algorithms can write texts, device marketing campaigns, etc., the current version of the question is: where, then, are humans (“in-the-loop”) still undispensable, what is the new core of publishing, how can the system of publishing be described in the age of AI? This contribution will feed the discussion on what AI does to publishing and particularly to a convincing concept of publishing by discussing various thoughts on the essence of publishing and by postulating an irreducible core of publishing – irreducible not least in the sense that tasks in this core cannot be accomplished by algorithms in the general case or in a sufficient quality, respectively. Based on this, the contribution will propose a conceptualization of the “humans-in-the-loop” workflows around this core we have to expect and already see.