Man Brags About Using AI To Publish Children's Book In A Weekend, Does Not Receive Praise
Anyone who has been on Twitter for any amount of time knows that the one faction you absolutely do not mess with is Book Twitter — a group notoriously prone to controversies and infighting. However, one tech bro was able to unite the various sects against a common enemy by proudly declaring he'd used AI to bang out a children's book over the course of a weekend.
On Friday, Twitter user Ammaar Reshi posted a thread detailing how he'd used various AI tools to make Alice and Sparkle, a children's book that is, in essence, written and illustrated by artificial intelligence about a young girl learning about the magic of artificial intelligence.
The thread details how he fed writing prompts into ChatGPT (like "Describe what Sparkle looks like"), and the AI spit back paragraphs that he would then tweak with a new prompt. Reshi describes this process as "like having a constant brainstorming partner who I could ping pong ideas off of."
He then went to Midjourney and fed it prompts until he could get a consistent art style. He stitched the results together and published Alice and Sparkle through Amazon Kindle Publishing.
Considering Reshi did not write or illustrate anything in the book he will receive the profits from, his venture rubbed many the wrong way, and hardly elicited the "oohs" and "wows" his thread seemed to encourage. Instead, most tweeters were quite harsh in their criticism of Reshi's endeavor — as is often typical of AI art criticism nowadays.
Some discussed how children's books take imagination and creativity that is clearly beyond AI capabilities, while others took a more streamlined approach in their criticisms.
As for the book itself, the snippets that Reshi showcased did not appear to impress many annoyed Twitter users. One post by @Coreybrickley pointed out the AI's numerous basic illustration errors that would never have been submitted by a competent illustrator.
Reshi then responded to the controversy by noting that in response to his thread, there were many "people who are incredibly excited about empowering tech giving them the ability to create things they couldn’t before" (though such posts are difficult to find), that there are artists concerned about the implications his venture has for their livelihoods and "keyboard warriors" who are sadly not open-minded about the topic.
This last point was received about as well as you might expect.
Share Pin
SardonicRainboom
Well, you seem to have skipped the part where we discuss whether or not it qualifies as theft.
lecorbak
he doesn't care, he knew what he did and expected backlash so that more people buy his book, and people fell to this trap.
KoimanZX
Quality aside, this is the sort of thing that worries professional artists.
Gumshoe
I don't think you can really say "quality aside" though. The fact that the quality is so low is why this shouldn't really be that worrying, at least for now. If this is the alternative, then I just really don't see many parents wanting to buy this instead of a regular children's book.
Uboa
"It was a social experiment guys, I swear"
Revic
Philosophy and ethics aside: The technology ain't there yet, bro. AI can be used as a starting point but if you want something suitable for publication you're going to need to put in a lot more work on top of that. It's cool, impressive technology but it's not going to carry someone who doesn't want to do any work.
Gumshoe
Yeah I think the same. This kind of thing seems useful to help brainstorm ideas or get a rough starting sketch or whatever, but you're still going to need some art and writing skills to make it into something people will actually want to buy. I don't know art but I know coding, and people had similar reactions there. The reality is that Chat GPT code very often has pretty bad mistakes in it, which is fine if you are skilled enough to recognise and fix them before putting it to use, and even then it's only good for very basic problems that have pretty generic solutions.