Will Paying Artists Win Them Over to AI?

Illustrators have spent years raising alarms about generative AI companies training their models on artists’ work without permission. They’ve called the practice outright theft, and in response, many AI advocates have insisted it’s a necessary step for the technology to advance. That clash has sparked heated legal fights, but it’s also opened the door for a new generation of startups like Pippa, which markets itself as a more ethical alternative in the crowded field.

Pippa’s core offering mirrors most text-to-video platforms: short, user-generated clips that people can stitch together into visual narratives. What actually separates it from the pack is how its business model routes money back to artists. Every time a subscriber generates an image or video clip that draws on a human artist’s style, that artist gets paid by the platform. It’s a straightforward concept, but one that few competitors have embraced.

Co-founders Hogan Shrum and Sean Wright believe this artist-first approach will win over customers. They’re also betting that Pippa can demonstrate how AI companies can operate with a clear conscience. But the road ahead is steep. Even with its ethical framing, Pippa is still launching with the same rough-around-the-edges models that define other text-to-video services, and convincing skeptical artists to sign on won’t happen overnight.

In a recent conversation, Shrum and Wright told me they view Pippa as a way to move past what they called “the bloody history that the AI industry has been built on.” Their goal is to set a new standard for peers and show the public that AI tools don’t have to rely on creative theft to function.

“All of these models from the past 18 months have been trained on real people’s art without their permission, that’s how it was built,” Wright said. “I harken it back to when Napster was a thing and everybody was stealing music. But then, Apple came out with the 99-cent song, Napster got shut down, and we figured out a different way to get musical artists paid.”

The payments Pippa sends to artists are modest on their own, especially when stacked against the platform’s subscription fees, which run from $14.99 to $99.99 per month. Artists earn $0.005 per image and $0.003 per second of video generated from their style. As users create more content, those payouts climb, and artists also get a slice of a 5 percent royalty pool funded by subscription revenue. In its promotional materials, Pippa compares this structure to Spotify’s, which is a curious reference given how many musicians have publicly criticized the streamer’s compensation model.

Pippa’s technology still leans on art scraped from the internet

Pippa launched in May and currently counts around 800 paying subscribers. The company has signed licensing and model training agreements with just four human artists so far, though Shrum says negotiations are underway with four more. Interested artists must pass a multi-step vetting process to prove the illustrations they submit are genuinely theirs. Once approved, models trained on their work become available to subscribers, and each artist gets a profile page to direct fans to their broader portfolio. Wright also noted that partner artists can choose to submit under pseudonyms if they’re worried about being seen as “crossing the picket line.”

“We’ve had so many conversations with artists where they’re like, ‘I love this, but I don’t want to get ostracized by my community,'” Wright explained. “We’re very early in this space, so there’s some apprehension to participate in it. But we think that can change as people see the economics of how we’re paying artists and that our intent is good.”

Despite its ambitions, Pippa hasn’t fully transitioned to models trained exclusively on its in-house artists’ data. The company admits its current technology is built on open models that “have initial training on the broader set of content out there.” In plain terms, Pippa’s system still involves some degree of using art scraped from the internet without creators’ explicit consent. The intentions may be noble, but there’s no clean solution to the creative theft baked into generative AI.

This isn’t a problem unique to Pippa. Every “ethical” AI startup that relies on pre-existing, non-proprietary data faces the same dilemma. That’s why some companies, like Ben Affleck’s InterPositive, emphasize building wholly original datasets from scratch. But that approach demands significant upfront capital that most startups lack, and the end result is often highly specialized tools that don’t appeal to everyday users.

Browsing Pippa’s online portal, where users can watch others’ AI animations or generate their own, I didn’t see much that stood out from other text-to-video services. The site is currently dominated by kid-friendly content that struggles to mimic Pixar’s polished look, and very little appears inspired by an independent artist’s distinctive hand.

That’s somewhat expected since Pippa still relies on many of the same cloud-based models other platforms sell access to. In the near future, the company plans to integrate ByteDance’s Seedance 2.5 model, which could draw subscribers thanks to its ability to generate up to 30 seconds of footage that can be refined without restarting. But to truly differentiate itself from competitors offering the same model, Pippa will need to onboard more in-house artists, and that hinges on how well it can pitch itself to potential partners.

Shrum and Wright are confident more artists will come around, and that using Pippa will convince them to bring their communities along. For that to happen, though, public sentiment about generative AI would need to swing back in the technology’s favor, and that’s far from guaranteed.