Just about everyone has made doodles in MS Paint—or at least something equivalent. There’s something charming about the way that simple lines can create something recognizable to the human eye, and something impressive about someone using the app to make a genuinely expressive piece of art. Now, what if we removed all of that charm and just jacked the style, without any of the human touch that makes it interesting in the first place? Well, then you’d get the latest ChatGPT trend.
Users have started prompting OpenAI’s latest image generation model, GPT Image2, to create MS Paint-style drawings of their photos. The results are about what you’d expect, really. They are recreations of existing images in simple, squiggly lines with minimal color that are clear enough to be recognizable. Cool.
The trend appears to have started, or at least been popularized, by a Korean user on Threads with the handle @withgrdnrush. They shared several doodle-style images created with GPT Image2, calling it “The Most Trivial Prompt in the World,” per a translation. They didn’t include the original images, but the outputs look like something an artist could throw together in MS Paint in a matter of minutes.
The replies quickly filled up with other people making use of the prompt, and the trend has since spread, making its way over to the AI-obsessed folks on X. A user with the handle @arrakis_ai, who seems to mostly just aggregate AI content, posted about the “trend,” and included a prompt: “Redraw the attached image in the most clumsy, scribbly, and utterly pathetic way possible. Use a white background, and make it look like it was drawn in MS Paint with a mouse. It should be vaguely similar but also not really, kind of matching but also off in a confusing, awkward way, with that low-quality pixel-by-pixel feel that really emphasizes how ridiculously bad it is. Actually, you know what, whatever, just draw it however you want.”
The output doesn’t really seem to accomplish that. It definitely looks like an MS Paint drawing, but it’s not really off in any way, which is fitting that the machine would strip out any of the human charm of the whole thing.
@arrakis_ai suggested the trend could be “the Ghibli controversy, Season 2.” That seems extremely unlikely for a few reasons. First, at the center of the “Ghiblify” situation was the fact that the prompt to turn photos into Studio Ghibli-style drawings appeared to infringe upon a trademark style. Whether you believe GPT Image captured the nuances of that style or not, it takes some skill to approximate it. Second, OpenAI CEO Sam Altman was the one who really spread that trend, encouraging people to feed their images into ChatGPT to get the Miyazaki-knockoff outputs.
There is no indication that the MS Paint prompt has anywhere near the virality, nor is there any reason to think that it is in violation of any copyrighted material or styles. It’s just a silly prompt that creates a recognizable output while eating up considerably more resources than necessary. The original version of MS Paint ran on PCs with 8-bit CPUs and 512KB of RAM. Now people are tapping into gigawatt data centers to get the same final product, without any of the creativity. We’ve come a long way.
Read the full article here
