Readers of this blog will know that I regard Luciano Floridi as one of the more insightful guides to the conceptual issues of a rapidly changing information environment, and his philosophy of information as a valuable basis for library/information science. In an open access editorial in Philosophy and Technology (31st August 2024), Floridi explores the future of content in the age of generative AI (ChatGPT and the like), analysing the evolving definition of content, the transformations brought about by GenAI systems, and emerging models of content production and dissemination. His contention is that, as GenAI becomes more sophisticated, and distinguishing between human- and AI-produced content becomes increasingly difficult, the trend towards a blurring of the traditional boundaries between content producers and consumers has accelerated this trend, challenging our traditional understanding of content and opens up new possibilities for creation and consumption.
Floridi reminds us that content, in his usage, is comprised of well-formed and meaningful data. Data is Floridi’s philosophy of information amounts to a lack of uniformity, a difference, The data which makes up content may be analogue or digital differences, encoded in some consistent form, such as alphabetic symbols or binary codes. The structure of the data, their syntax, determines their well-formedness, while their meaning constitutes their semantics. Random data, without syntax, or meaningless data, without semantics, could be considered content, but this is not the norm. Content usually means structured, meaningful data, made available for purposes such as entertainment, education, or communication; Floridi’s examples are blogs, news items, social media posts, pictures, podcasts, video clips, or interactive computer games. Floridi regards content, understood this way, as equivalent to information in a weak sense; a stronger sense requires the content to be true to count as information. (Parenthetically, I have found this concept of semantic content to be very helpful as a way of conceptualising information for the library/information sciences. It gets us away from the vague and disputable ideas that libraries, databases, etc. contain information or knowledge.)
Using this idea, Floridi examines three classic models through which content is produced, disseminated, and consumed: the input-process-output (IPO) model, the Shannon’s model of communication, and the supply-demand model. With this analysis he shows that the impact of the move to a world in which a significant amount of content is produced by GenAI is “far-reaching and profound”. To take one simple example: “if authors can produce whole books using GenAI, so can readers, choosing according to their preferences. Therefore, in the future, Amazon may move from restricting authors from publishing too many A I-generated books to enabling readers to have AI write the books they wish to read, à la carte”. This transition may also lead to the rise of highly personalized content. This will differ from the “Daily Me” idea of the 1990s, in that the Daily Me implied a consumer-driven selection and presentation of content, whereas GenAI will facilitate consumer-driven production. Didn’t like way that TV series ended? Get GenAI to create new series finale more to your taste. Wish the author had written another book in that series? Get GenAI to write one, with an ending to your liking. [My examples, not Floridi’s.]
“We will”, says Floridi, “probably witness the emergence of new content cultures, where personalised content experiences may give rise to new forms of cultural exchange and identity formation. These changes may reshape our understanding of culture itself, with the history of culture potentially being reframed (also) as the history of content production, dissemination, curation, fruition, and transmission.” Very probably. And it is easy to imagine the issues which this will raise for the information and documentation sciences and professions; the nature and status of such documents, the decisions as to what should be collected, and the forms of metadata necessary to describe and retrieve such content, being only the most obvious. Challenging times indeed, and we will need conceptual (yes, philosophical) analyses such as these to guide research and practice in coping with them.
The Occasional Informationist 