Amazon: Machine-Generated Content Adds to Overhead Costs
July 7, 2023
![Vea4_thumb_thumb_thumb_thumb_thumb_t[1]](http://arnoldit.com/wordpress/wp-content/uploads/2023/07/Vea4_thumb_thumb_thumb_thumb_thumb_t1-10.gif "Vea4_thumb_thumb_thumb_thumb_thumb_t[1]")
Note: This essay is the work of a real and still-alive dinobaby. No smart software involved, just a dumb humanoid.
“Amazon Has a Big Problem As AI-Generated Books Flood Kindle Unlimited” makes it clear that Amazon is going to have to re-think how it runs its self-publishing operation and figure out how to deal with machine-generated books from “respected” publishers.
The author of the article is expressing concern about ChatGPT-type outputs being assembled into electronic books. That concern is focused on Amazon and its ageing, arthritic Kindle eBook business. With voice to text tools, I suppose one should think about Audible audiobooks spit out by text-to-voice. The culprit, however, may be Amazon itself. Paying a person read a book for seven hours, not screw up, and making sure the sound is acceptable when the reader has a stuffed nose can be pricey.
A senior Amazon executive thinks to herself, “How can I fix this fake content stuff? I should really update my LinkedIn profile too.’ Will the lucky executive charged with fixing the problem identified in the article be allowed to eliminate revenue? Yep, get going on the LinkedIn profile first. Tackle the fake stuff later.
The write up points out:
the mass uploading of AI-generated books could be used to facilitate click-farming, where ‘bots’ click through a book automatically, generating royalties from Amazon Kindle Unlimited, which pays authors by the amount of pages that are read in an eBook.
And what’s Amazon doing about this quasi-fake content? The article reports:
It [Amazon] didn’t explicitly state that it was making an effort specifically to address the apparent spam-like persistent uploading of nonsensical and incoherent AI-generated books.
Then, the article raises the issues of “quality” and “authenticity.” I am not sure what these two glory words mean. My impression is that a machine-generated book is not as good as one crafted by a subject matter expert or motivated human author. If I am right, the editors at TechRadar are apparently oblivious to the idea of using XML structure content and a MarkLogic-type tool to slice-and-dice content. Then the components are assembled into a reference book. I want to point out that this method has been in use by professional publishers for a number of years. Because I signed a confidentiality agreement, I am not able to identify this outfit. But I still recall the buzz of excitement that rippled through one officer meeting at this outfit when those listening to a presentation realized [a] Humanoids could be terminated and a reduced staff could produce more books and [b] the guts of the technology was a database, a technology mostly understood by those with a few technical conferences under their belt. Yippy! No one had to learn anything. Just calculate the financial benefit of dumping humans and figuring out how to expense the contractors who could format content from a hovel in a Myanmar-type of low-cost location. At night, the executives dreamed about their bonuses for hitting their financial targets and how to start RIF’ing editorial staff, subject matter experts, and assorted specialists who doodled with front matter, footnotes, and fonts.
Net net: There is no fix. The write up illustrates the lack of understanding about how large sections of the information industry uses technology and the established procedures for dealing with cost-saving opportunity. Quality means more revenue from decisions. Authenticity is a marketing job. Amazon has a content problem and has to gear up its tools and business procedures to cope with machine-generated content whether in product reviews and eBooks.
Stephen E Arnold, July 7, 2023
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
