When I studied literature we mostly read the classics. Great literature, the canon. But that’s not necessarily what most people actually read. What if instead of comparing AI-generated literature to the literary canon, we tried comparing it to super popular and commercial forms of literature instead? Like the folkebøker that were popular in Norway from the 16th century on, or the market literature sold in African markets?

Anne Sigrid Refsum joined us this week at the Center for Digital Narrative as the first postdoc to join the AI STORIES project team. On Tuesday, the day after she officially started working at the CDN, an article she wrote on folkebøker or popular books in Norway from the 1500s on was published in the journal Tidsskrift for kulturforskning. Anne Sigrid’s PhD project was about skillingsviser or shilling ballads, which are popular songs sold and sung for almost 400 years. (Here is the project she was part of.) Her application for the postdoc on AI STORIES was brilliant: she argued that her expertise on older forms of popular, commercial literature gave her the perfect background to analyse AI-generated stories. I was convinced, and reading this new article I think this will be a super-useful background.

In France they were known as biblioteque blue because they were printed with blue covers. In Norway they were called folkebøker and in Germany Volksbücher, and they were extremely popular. Salespeople walked around with baskets full of printed song lyrics and these chapbooks with stories about love, magic and knights, and they were read and reread throughout Europe. Anne Sigrid mentions novel factories in Germany in the 16th and 17th centuries, and as soon as I read that I started imagining what a novel factory would look like: rows of writers at desks, and of course lots of printing machines. Then I thought it might be an anachronism – didn’t factories come later? But no, in fact the Oxford English Dictionary’s first reference for the word factory in the modern sense is about a book factory!

Yesterday Kwabena Opoku-Agyemang gave a talk on The Rhetoric of Space in African Digital Literature here at the Center for Digital Narrative, and in the Q&A session afterwards he mentioned African market literature, which he described as a long-established genre of extremely popular stories that are sold on cheap paper in markets and that are not recognised as “real” literature.

Kwabena talked about short stories published in online African literary journals like Jalada, Flash Fiction Ghana, Saraba, Adda, Brittle Paper and African Writer Magazine, discussing how their use of place and setting played to stereotypes about Africa but also challenged the stereotypes. It was an interesting talk; we recorded it and plan to publish it on the CDN YouTube channel when we have time to process the video.

Market literature in Africa describes very cheap and not very well-written stories that are sold in markets and stations where you would get a mini bus, so these are very transitional spaces, Kwabena explains. The spelling is often off and they are poorly printed: these are books people read just to pass the time. The writing is both awful and glorious: always a guy meets a girl, there is often a spiritual attack. The website EbonyStory contains many such stories. Here is an article by Emmanuel Obiechina summarising the genre of market literature in Africa.

There are a lot of explicitly AI-generated African stories online as well. AI African Stories is a YouTube channel that a previous guest to CDN, Tolulope Oke, showed us. Here’s a video explaining how to get rich from AI-generating stories.

I think Anne Sigrid is on to something: AI-generated literature is often more akin to these types of popular literature than to the literary fiction critics usually focus on. Perhaps it would be useful to stop calling it “slop” and instead thinking of how it relates to these popular forms, which a lot of people clearly actually want to read.