Lucas Freitas's AI-assisted art depicting countries as villains

Originally published at: Lucas Freitas's AI-assisted art depicting countries as villains | Boing Boing

1 Like

Yawn. Don’t care.

I hope we get bored of these “Someone asked AI to depict ABC as XYZ!” stories sooner rather than later.

Or find the plethora of artists who have done cool series like this and highlight them.


Well, in the parlance of Hollywood, more like story credit and art direction by human…


I’d like it a lot more if y’all would spotlight some real artists.


Unlike the folks above, I’m actually interested… but since they posted it to Facebook Jr. and I don’t have an account, and refuse to get one, I can’t even see it. Maybe stop bothering to post stuff if it’s only on Instagram, or any other site that requires a login to access content?


I am not sure what is going wrong for you, but I can see it just fine without any kind of login.


There are two types of AI art I will support:

  1. An algorithm that’s trained on public domain art, creative commons art, and/or art where the artists themselves have given permission for their art to be included in the training set. Such a program should be available to anyone to use, and no art - NONE - from any rights owner that has not opted in should be used.

  2. AI art that is controlled by a single artist, trained entirely on their own body of work, giving them the ability to rapidly prototype and develop new twists on their own work.

This work, though I love the concept and some of them are wonderful, is a ginormous rip off of other people’s work, derivative, and likely required almost no fucking effort of the “artist” in question beyond coming up with some random text to feed into it. But it sure as hell is making a TON of fucking money for the developers of Midjourney because people have to pay after the initial few free runs they can use.

This is tech bro art. It is built on stolen work, and designed to make money for the tech bros, not actual artists. And government won’t do jack shit until it’s far too late to stop.


I looked; as per uz, it’s not as bad as the naysayers would lead one to believe:


I can only assume this one is the US, but the heading of the post seems to (intentionally?) obscure the text of the image.

Without seeing the actual resources and input used to generate each image, I can’t say what is “ripped off” and what is inspired by: what I can say is that it’s totally possibly to use AI to remix existing art which hasnt been appropriated or stolen, to create something new.

For instance:

That’s my own art, run through a free* online generator.

*It was free when I last used it anyway…


Your assumption is correct. If you follow the link to the Instagram account you can see more collections of these photos (including one where the USA title is unobscured) as well as photos for additional countries such as Finland, Argentina, Ireland, New Zealand and others. But I don’t have an Instagram account so it only let me view a limited number before asking me to sign in.

As one might expect (whether the art is generated by biased AI or just regular biased humans) most of them tend to be based on whatever stereotype tends to come to mind for people who only have a passing familiarity with a country. So of course Ireland has an evil leprechaun, Finland has a Norse Viking looking guy, etc. Some colonized countries have their “villain” in the form of the indigenous people though, which is maybe a plus for diversity (Māori warriors are obviously cool) but still definitely problematic in this context, where the indigenous folks weren’t the oppressors. As the first humans on the island the Māori weren’t in a position to oppress anyone other than maybe the Moa.

(Edited to clarify that the second paragraph was responding to @dnealy)

1 Like

Wales puzzles me

1 Like

Yeah I mean why not a big red dragon?




A cleverer AI might have picked up on the Mari Lwyd

1 Like

This topic was automatically closed after 5 days. New replies are no longer allowed.