Pippa pays artists per AI image—but still uses scraped art underneath

Generative AI startups have long faced criticism from illustrators and other visual creators for training models on artwork harvested from the internet without permission. Pippa, which launched in May, represents a new breed of company attempting to address these concerns through direct artist compensation.

The platform operates as a text-to-video service where subscribers pay monthly fees between $14.99 and $99.99 to generate short video clips and images. What distinguishes Pippa is its commitment to paying artists whenever their style is used. According to the company, artists earn $0.005 for each generated image and $0.003 for each second of generated video. Beyond per-use payments, artists receive a share of a 5 percent royalty pool derived from subscription revenue—a structure Pippa compares to Spotify's model, though the music industry's own frustrations with streaming economics make this an awkward reference.

Cofounders Hogan Shrum and Sean Wright see Pippa as a way to move past "the bloody history that the AI industry has been built on." In an interview, Wright drew a parallel to the Napster controversy: "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. 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 duo argue that viable, paid pathways for creators could shift industry norms.

Pippa's onboarding process requires potential partner artists to undergo a multi-step vetting process to confirm ownership of their work. Once approved, an artist's style becomes available to subscribers, and artists receive a profile page to promote their broader portfolio. The company has signed agreements with four artists and is in talks with four more. To address concerns about community backlash, Pippa allows artists to participate under pseudonyms. "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.

However, a significant caveat undermines Pippa's ethical pitch: the platform still uses AI models with training on broadly scraped internet content. The company intends to transition to models trained entirely on datasets from its in-house partner artists, but that infrastructure is not yet operational. Instead, Pippa relies on open-source models—including plans to integrate ByteDance's Seedance 2.5, which can generate up to 30 seconds of video with fine-tuning capability—that inherently include unlicensed creative work in their training data. As the article notes, "though Pippa hopes to fully switch to models that are trained on datasets provided by its in-house artists, that technology is still built with open models that the company says 'have initial training on the broader set of content out there.'" This means Pippa has not solved the fundamental problem of creative theft baked into current generative AI; it has only improved incentives for participating artists.

With around 800 paying subscribers as of the article's writing, Pippa is still in early stages. Its competitive position depends on signing more artists whose distinctive styles can differentiate it from competitors also offering the same cloud-based models. Success will ultimately rest on whether artists view the payment terms as worth the reputational risk and whether the broader creative community begins to accept generative AI as a tool rather than a threat.