Procreate is drawing a line in the sand—it will not build generative AI into its app, nor will it collect user data for AI training purposes. The company explains that generative AI, which is “built on a foundation of theft,” impedes on human creativity and should not be utilized by an app that “deeply respects” visual artists.
This ethical and legal stance, which was relayed by Procreate over the weekend, appeals to a large segment of the visual art community. It’s also an interesting business move that helps Procreate differentiate itself from other raster art software, including Adobe Photoshop, which has effectively gone all-in on AI.
The statement that AI-generated images equate to “theft” is particularly notable, as it echoes a popular sentiment among visual artists. Most AI image generators are trained on unlicensed photographs, digital paintings, and illustrations. Under normal circumstances, a company that uses unlicensed art for commercial gain would be slapped with a copyright claim from the artist. But artists have no way of knowing when their work is utilized for AI training, so filing a copyright claim or seeking compensation is practically impossible. World governments have also failed to address this issue in any substantial way, corporations reject that generative AI is theft (until they themselves are made the victim), and there is no legal precedent to incentivize individual or class action lawsuits from victimized artists.
“Generative AI is ripping the humanity out of things. Built on a foundation of theft, the technology is steering us toward a barren future. We think machine learning is a compelling technology with a lot of merit, but the path generative AI is on is wrong for us … We’re not chasing a technology that is a moral threat to our greatest jewel: human creativity.”
Most companies that embrace generative AI are keen to acknowledge the legal and ethical dilemmas presented by Procreate. For its part, Adobe regularly argues that its AI products are “commercially safe” and in full compliance with copyright law. The company follows strict in-house ethics guidelines and trains its models on licensed Adobe Stock images, rather than random images found on the web.
But, even with these guardrails, Adobe’s situation is pretty murky. We recently learned that Adobe Firefly’s data set contains images from Midjourney, an AI model that’s fueled by unlicensed material. Customers have also expressed concern that the artistic projects they upload to Creative Cloud may be covertly funneled into AI datasets (Adobe denies this claim). Understandably, the opaque and somewhat deceptive nature of Adobe’s AI ambitions (which is reflected in dozens of other corporations) has not alleviated the concerns of those who question or criticize generative AI.
Procreate’s anti-AI stance is sure to draw encouragement from many artists, which may be good for the company in the short term. The question is whether this stance can be maintained over the coming years. If a rival digital painting app develops useful (and ethically decent) generative AI functionality, Procreate could lose market share. The only reasonable response to this situation, at least from a business standpoint, would be to backtrack on the anti-AI promise.
This is all hypothetical, of course. Procreate has gotten on fine without generative AI. And because Procreate is designed for digital illustration or painting, rather than photo editing, the potential benefit of generative AI is marginal (at least with current technology). The few artists who want to incorporate AI into their Procreate workflow almost certainly need a detailed photo editor, too—these artists already dump their Procreate projects into Photoshop, and they would continue doing so if Procreate took up some AI functionality.
