In reply to Stop Using AI-Generated Images by Robin Rendle

This specific argument sounds way too much like the ‘every music download is a missed sale‘ trope of the music industry, 2-3 decades ago. And that was about actual recordings, not generated ones.

There are many good reasons to not use algogens, or to opt for different models for such generation than the most popular public facing tools.

Missed income for illustrators by using them in blog posts as Rendle posits isn’t a likely one.
Like with music downloads there’s a whole world of potential users underneath the Cosean floor. Several billion people, I assume.

My blog or presentations will never use bought illustrations.
I started making lots of digital photos for that reason way back in 2003, and have been using open Creative Commons licenses whenever I used something by another creator since they existed.
These days I may try to generate a few images, if it’s not too work intensive. Which it quickly is (getting a good result is hard and takes more time than a Flickr search for an image) and then paying someone who can manipulate the tools better than I (like now with Adobe Illustrator) might be in order, except for my first point at the start of this paragraph. Let’s also not pretend that all current illustrators do everything by hand or have done since the 1990’s. There’s tons of art work out there that is only possible because of digital tools. Algogens aren’t such a big jump that you get to treat them as a different species.

Not to say that outside the mentioned use case of blogs and other sites (the ones that already now are indistinguishable from generated texts and only have generating ad eyeballs as purpose), the lower end of the existing market won’t get eroded unless the lower end of the market ups their game with these tools (they’re actually better positioned for it than me in terms of skills and capabilities).
I bet that at the same time there will be a growing market for clearly human made artefacts as status symbol too. The Reverse Turing effect in play, as the default assumption will be, must be, that anything is generated. I’ve paid more for prints of artwork, both graphics and photos, made by or at the behest of the artist, than one printed after their death for instance. Such images adorn the walls at home rather than my blog though.

I always thought those AI-generated images in blog posts and elsewhere around the web were tacky and lame, but I hadn’t seriously considered the future economic impact on illustrators.

Robin Rendle

Image generated with Stable Diffusion using the prompt “A woman is sitting behind a computer making an image of a woman drawing, sketchbook style”. The result is an uncanny image. Public domain image.