A good point, made in a new short OA opinion-piece by Martin Eve…
How can the humanities parrot the oft-repeated liberal humanist line that they exist to produce an educated citizenry capable of participating critically in democracy, when most humanities work remains unreadable by most people? … When justifying oneself, pointing at a scholarly book that costs £60 is not the same as pointing to an article that can be read for free.
Assuming that ‘democracy’ and ‘educated citizenry’ are not soon to raise a sneer rather than a cheer from those holding the cattle-prods, then an obvious solution might be some form of free digest of the text that one wishes to read or consult. Often a good digest might suffice to quench curiosity on the part of independent scholars outside salaried academia. It would be far more than an abstract or a paragraph in an editor’s introduction, better than an Amazon Kindle-like “the first 10% is free” extract, and yet shorter than a Blinkist full-book summary (these being carefully crafted by a human, last I heard).
Could a summarising AI do the task? It can with newspaper articles and (I believe) with structured things such as financial reports. It might need some help with denser and more specialised material from the arts and humanities. Such an AI-bot might be aided by picking up on a dozen simple ‘structure tags’, added by the author alongside the text as they wrote it. And in that task the author might themselves be being assisted by a tagging AI. There might also be a semantics back-end at work. A local history chapter on well-dressing ceremonies and associated folklore in the English Peak District might then not flummox the AI too much. But a convoluted chapter on Elvish linguistics and arcane medieval star-lore detectable in Tolkien’s The Silmarillion might still cause it dissolve into a hissing molten puddle.
What about the idea of digests written by young humans in need of cash, of which the planet is very soon to have a rather large abundance? The AI digester-bot might then assist a well-trained young human, by providing a ‘first pass’ digest.
An informed and curious citizen might then be given 52 credits a year to access such a digesting service. Each credit would pay for the cogent summary of a book chapter or article. The specialisation and reading-difficulty of a text would be assessed initially by the AI, with some texts deemed to merit the expenditure of more than one credit. The digestion of a dense book which evaluates the Norse linguistics allegedly to be found in Sir Gawain & The Green Knight might even need a crowd-funding consortium or similar, pooling their collective credits. But that seems unlikely.
My idea loses ground again when one considers it might work poorly with summaries-of-summaries. For instance I want the weighty “The Year’s Work in Tolkien” summary overview in the latest Tolkien Studies journal. But the issue is prohibitively expensive in print, and is locked away from the public on Project Muse. As an impoverished independent scholar I need to read every word of it to keep up with the field, and a summary will not suffice. No matter how good it is.
So… none of this is really ideal, though it would certainly create a welcome market in somewhere like Bangladesh for AI-assisted ‘academic summarisers’ servicing a ’52 credits a year’ system.
One other vague notion also arises. Google Books already exists and provides a partial ‘Look Inside’ solution for many who need to take a peep into an expensive £50-£120 academic collection or (far less often available) an obscure monograph. Could that existing service be expanded? Even if only through gritted teeth, as most librarians seem to despise Google Books. How about a mandate that says Google Books gets to show 100% of a volume produced with or originating from a public university employee, but only ten years after publication? Could that work in terms of the current economics of such things? I don’t know enough about publishing’s current ‘profits over time’ aspects there, as I haven’t been following the monographs debate. But such a book would still be effectively locked down (not OA in any meaningful sense), while still being readable by the global public. It would be a sort of automatic ‘Knowledge Unlatched’, running globally alongside the existing copyright systems and (because universal) not subject to political skew in terms of the books selected. It might also be retrospective.