Artists decry use of AI-generated art: ‘I’m concerned for the future of human creativity’

New tech raises labour, privacy, and gender issues

In recent weeks, artificial intelligence apps have gone viral on social media for allowing users to create avatars in the style of various famous artists. These powerful new tools are changing more than just people’s profile pictures, however; according to artists and creatives, they may be changing the face of creative labour in permanent, frightening ways, all while raising serious privacy and intellectual property concerns.

AI-generated art is suddenly everywhere you look.

Apps like the photo editor Lensa allow users to create “magic avatars” in a near-infinite variety of genres. It’s been a huge hit with users: since Lensa launched the avatar feature in November, over 4 million people have downloaded the app, spending $8m on internal features, according to WIRED.

But it’s not just photos. OpenAI’s GPT-3 can produce pieces of eerily human-like writing based on text prompts from users.

Major companies like Microsoft and Adobe are also integrating AI tools into their offerings.

The prospect of easily accessible tools that can closely approximate human artistic output has many creatives worried.

“I’m incredibly anxious for the future of my career, more than ever before,” artist Kelly McKernan wrote on Twitter. “Further, I’m concerned for the future of human creativity.”

The art of Ms McKernan, a painter and illustrator with a cosmic, surreal style, was one of the early tranches of images used to train Stable Diffusion, a popular tool used in AI art apps.

In a thread, the artist described how “at first it was exciting and surreal” to help inform an AI studying the building blocks of creativity, but later was a trip through the “uncanny valley” when Stable Diffusion users began spitting out close imitations of her work en masse.

What’s more, some of these users began taking images clearly based on Ms McKernan’s work and using it for their own purposes, commercial and otherwise, balking when she would request her name be removed from tagged images in her style.

“Please don’t support the unethical use of AI image generators while thousands of artists are infringed upon,” she concluded. “Demand better, and please keep speaking out! If artists can’t defend the use of their names and artwork, what have we got left?”

Beyond general labour concerns, many in creative fields accuse AI of violating their intellectual property.

AI models like Stable Diffusion, the basis for Lensa’s magic avatars and other tools, use huge caches of publicly available images to train themselves about the nuances of different artistic styles.

As a result, these AI models harvest the stylistic DNA of individual artists, then allow strangers to borrow elements from their work without offering any credit. Further, since many AI models are prompt-based, sometimes this borrowing process is incredibly direct.

For example, nearly 100,000 Stable Diffusion users have prompts directly naming Greg Rutkowski, a fantasy illustrator who has worked on games like Dungeons & Dragons. The images they create are based on his work, but can be used for any purpose they want.

"We could say that, ethically, it’s stealing," Mr Rutkowski told the CBC.

Despite these concerns, AI is such novel territory in the legal world that it’s unclear how an artist like Mr Rutkowski could protect his IP from being sucked into AI models even if he tried.

“I see people on both sides of this extremely confident in their positions, but the reality is nobody knows,” technologist Andy Baio told The Verge. “And anyone who says they know confidently how this will play out in court is wrong.”

Other critics point out how apps like Lensa, trained off what’s essentially a sampling of the whole Internet, amplify the misogyny and predatory aspects of some corners of the web.

Some users report AI image generators spitting out highly sexualized photos, including nude pictures, when fed innocuous selfies and childhood photos.

Prisma Labs, the company behind Lensa, has defended its app and products like it.

“AI produces unique images based on the principles derived from data, but it can’t ideate and imagine things on its own,” the company wrote in a Twitter thread. “As cinema didn’t kill theater and accounting software hasn’t eradicated the profession, AI won’t replace artists but can become a great assisting tool.”

“We also believe that the growing accessibility of AI-powered tools would only make man-made art in its creative excellence more valued and appreciated, since any industrialization brings more value to handcrafted works,” the company added.

Indeed, some in creative professions have argued AI is a help, not a threat, allowing them quick and cheap ways to generate professional-quality imagery.

“I think there’s an element of good design that requires the empathetic touch of a human,” Sabella Orsi, a San Francisco-based interior designer, told The New York Times. “So I don’t feel like it will take my job away. Somebody has to discern between the different renderings, and at the end of the day, I think that needs a designer.”