This Artist Uses AI to Imagine Famous Characters In Different Worlds And Settings, And It’s The Entertainment We Never Knew We Needed (45 Pics)

Digital artist Milan Jaram has revealed his method for bringing his thoughts to life using artificial intelligence. In an interview with Bored Panda, Jaram explained, "Random ideas in my head spark my curiosity. I plug all of them into AI to see what they would look like, then share the results."

While many of the prompts he feeds to AI don't turn out well, Jaram still enjoys the process. "Golden Girls as Mortal Kombat characters? I mean, why not? I love to see the AI struggle with these kinds of requests, and some of the horrifying images it spits out cement that," he said.

His sources of inspiration are varied, drawing from both his childhood and adulthood. "I mostly go [with] things that gave me fond memories, like Golden Girls or Mario World, [or something from] later in life, like The Office, King of the Hill... Hmmm... What if the King of the Hill cast were superheroes? (A random thought at 3 am). AI allows me to explore that. And again, share it with everyone else if I got a laugh out of it," he added.

Talking about the future merge of AI and traditional art forms, Jaram believes it will begin with concept art, storyboarding, and generating ideas. "[This] is a quick way to get some images out. Particularly for movie and game companies," he said.

"The artists can then work from those images as references. Now, depending on the employer or company, they might even want a simple 'paint over' if the AI got it right. Maybe AI can lay down the base image and artists can work on top of it. Some companies have a certain 'style' and AI can already achieve those styles, it's just a matter of constantly tweaking the prompts to get the image that is envisioned, and perhaps a paint-over by an artist."

Jaram thinks that smaller companies and developers who can't afford an artist will benefit from this the most. "They [will probably] get some art even with a small budget and potentially give us some games or visual novels that we otherwise wouldn't have seen."

However, some are criticizing the moral and legal aspects of the craft. For example, a first-of-its-kind US copyright lawsuit targeting AI art generators could limit the number of images the tools ingest for training, ultimately affecting the content that they produce.

A group of artists recently filed a potential class action against billion-dollar company Stability AI Ltd. along with two other art generator makers, Midjourney Inc. and DeviantArt Inc., over their use of copyrighted images to train AI tools. They claim that the generators downloaded and used billions of copyrighted images without obtaining the consent of or compensating any of the artists.

In addition to damages, the lawsuit — filed by Sarah Andersen, author of the webcomic Sarah Scribbles, and fellow artists Kelly McKernan and Karla Ortiz — asks the court to stop the AI generator companies from using artists' work without permission.

"It will change how the model performs if instead of billions of images, they're using much more tightly curated data sets," Ryan Abbott, an attorney at Brown Neri Smith Khan LLP, told Bloomberg. Though this lawsuit is the first involving AI-training images, it's reminiscent of the two programmers' November 2022 copyright suit accusing OpenAI, Microsoft Corp., and GitHub Inc. of using their code to train the Copilot AI coding tool without proper licensing.

Furthermore, Getty Images just announced that it initiated copyright infringement legal proceedings against Stability AI in a UK court, alleging it used Getty's digital images without a license.

Some attorneys agree that the copyright infringement allegations have merit. But Stability AI told Bloomberg in a statement, "Anyone that believes that this isn't fair use does not understand the technology and misunderstands the law." The success of a fair use defense will depend on whether the works generated by the AI are considered transformative—whether they use the copyrighted works in a way that significantly varies from the originals.

"The minute you’re able to make a work that’s transformative, you removed it from the ambit of this idea that it’s a mere derivative of the original work," Vivek Jayaram, founder of Jayaram Law, said.

Jayaram pointed out that previous case law, particularly the Supreme Court’s 2021 Google v. Oracle decision, suggests that using collected data to create new works can be transformative. Back then, Google’s use of portions of Java SE code to create its Android operating system was found to be fair use.

"AI doesn't 'steal' anything," Milan Jaram said. "It creates based on the input it is given. It can, however, copy styles. So yes, you can ask it for an image to be produced 'in the style of Bob Smith', and it will try its hardest to imagine the prompt you gave it, in that style. [But] that isn't 'stealing' from artists."

"We wouldn't have the myriad of awesome movies, games, music, architecture, art, fashion, and so on if it wasn't for people copying each other. And if you're an artist who actually sells art — both AI and artists can live together. Why? Because if someone wants to buy your art, they are buying it because they love it and maybe they love you or your style too. Art is subjective. Are AI images actually 'art'? That's completely subjective. Two people will never agree on what is or isn't art, and it isn't for you or me to dictate what is. If someone finds AI imagery as 'art', then good for them. If you don't, that's fine ... You can [see] a piece of bubble gum on the sidewalk as 'art', I might think you are crazy, but that decision is not up to me, it's up to the eye of the beholder."

Style has proven difficult, if not impossible, to protect with copyright.

"Style is an idea, style is a technique, it's a process," University of Kentucky law professor Michael Murray told Bloomberg. "These are things that are not protected by copyright."

Given an AI generator's ability to draw directly from the works of art it's trained on, however, there are circumstances in which "the line between emulating style and wrongful copying is not always apparent," Columbia Law Professor Jane Ginsburg added.

The circumstances may require a reckoning with current copyright standards.

Milan Jaram also highlighted that, "a very cool thing is that AI will use its best to give you things or styles that don't exist. For example, I did a TikTok with some prompts that included a 'scene from the movie Diablo 2 from 1985' ... and it gave me some scenes that looked exactly what you'd think it would look like."

"A Diablo 2 movie doesn't exist, but a game franchise does [and regardless] it did a very good job of putting all of that together." Here's a TikTok of a similar idea the artist tried out with The Legend of Zelda.