It is inspired by Midjourney’s new crappy automated moderation system that is always warning you about such and such innocuous thing being a “banned prompt.” Midjourney’s system uses two layers of AI, the second of which your prompt is sent to after you appeal the ban decision by the lower AI.
As someone who has spent years working on content moderation, I can confidently say that even humans can quite frequently not accurately detect whether a particular piece of content is indeed problematic. I don’t trust AI systems to do it AT ALL. And I think it’s not just a foolish & short-sighted direction that technology is taking, but that with the increasing integration of AI systems into all other technologies, it will prove in fact to be quite dangerous to humanity.
This book explores these concepts through the lens of short fiction, produced using Anthropic’s Claude, like most of the recent books in this series.
I’ll close this post with the image preview for the ebook: