Illustrators and visual artists have spent years raising the alarm about generative artificial intelligence startups training their models on creative work without permission. They argue that the practice amounts to theft, while many AI boosters counter that it is a necessary step for technological progress. This conflict has produced a wave of lawsuits, but it has also inspired a new generation of AI startups that claim to operate more ethically. One such company is Pippa, a text-to-video platform that hopes to prove revenue sharing can bring artists on board.
A different approach to AI video generation
Pippa is entering a crowded market where companies sell access to models that can produce short burst of video footage from text prompts. The service charges subscribers between $14.99 and $99.99 per month, allowing them to create visual stories through a simple interface. But what sets Pippa apart is its business model: every time a subscriber generates an image or video clip whose style is derived from a human artist's work, that artist receives direct compensation from the platform.
The company’s cofounders, Hogan Shrum and Sean Wright, believe this focus on paying artists will attract customers and set an example for the wider AI industry. In a recent interview, they described their project as a way to move past what they call “the bloody history that the AI industry has been built on.” They want to show that AI companies do not have to steal creative work to build successful products.
Wright drew a comparison to the music industry, recalling how Napster allowed widespread piracy before Apple introduced the 99-cent song. He argued that the AI sector is currently in its Napster phase, but a legitimate business model could emerge that treats artists better. “We figured out a different way to get musical artists paid,” he said. “We think we can do the same for visual artists.”
The payment structure: small amounts, big promises
Pippa's payments to artists are not large on an individual basis. The company pays $0.005 per generated image and $0.003 per second of video. A subscriber who creates a 10-second clip might generate just three cents for the artist whose style was used. However, as users generate more content, the payments add up. Pippa also distributes a portion of a 5 percent royalty pool funded by overall subscription revenue to its partner artists.
In its press materials, Pippa compares this payment model to Spotify’s. That comparison is somewhat ironic, given how many musicians have complained about the streaming service’s low royalty rates. Still, the company hopes that even modest payments will signal a change from the industry’s habit of taking without paying.
Artists who work with Pippa must go through a multi-step vetting process to verify they own the rights to the illustrations they submit. Once approved, models trained specifically on those illustrations become available to subscribers. The service also provides artist profile pages that can direct people to their work outside Pippa’s platform.
Wright noted that Pippa gives artists the option to submit work under pseudonyms if they worry about being ostracized by their communities. He acknowledged that many artists still harbor deep distrust of AI companies. “We’ve had so many conversations with artists where they say, ‘I love this, but I don’t want to get ostracized,’” he said. “We’re very early in this space, so there is some apprehension to participate.”
The uncomfortable reality of training data
Despite its ethical marketing, Pippa’s technology is not entirely free from the practices that artists have condemned. The company’s models are built on open-source foundation models that have been trained on large datasets scraped from the internet without explicit consent from every creator. Pippa says it plans to eventually shift to models trained only on its in-house artists’ work, but that day has not yet arrived.
This is not a problem unique to Pippa. Any “ethical” AI startup that relies on existing open models must reckon with the fact that those models were built using data that was often obtained without compensation or permission. Some companies, like Ben Affleck’s InterPositive, have made a point of developing wholly original, proprietary datasets from scratch. But building a custom dataset from the ground up requires significant capital and technical resources, which most startups do not have. The result is usually specialized AI tools that are not designed for mass-market consumers, unlike Pippa’s consumer-facing video service.
For now, Pippa is essentially a polished front-end for models that have already absorbed vast amounts of creative labor. Its internal models begin with “initial training on the broader set of content out there,” as the company puts it. That means even when Pippa pays artists for their contributions, the underlying technology still carries the fingerprints of thousands of creators who never had a say.
A confused product for now
A look at Pippa’s online portal reveals a service that is still finding its identity. The site is filled with user-generated animations that mostly imitate familiar children’s film aesthetics, often with limited success. Very little content on the platform appears to be inspired by an independent artist’s distinct hand. That is not surprising, given that Pippa currently uses many of the same cloud-based models that other platforms rent out.
The company says it plans to incorporate ByteDance’s Seedance 2.5 model into its service in the near future. That could attract subscribers because Seedance 2.5 can generate up to 30 seconds of video footage and offers fine-tuning capabilities without requiring users to start from scratch. But Seedance 2.5 is also available through other platforms, so Pippa will need to differentiate itself through its artist partnerships and payment structure rather than raw model performance.
To date, Pippa has only signed licensing and model training agreements with four human artists, though the cofounders say they are in talks with four more. The slow pace underscores the difficulty of convincing artists to trust an AI company, even one that offers compensation. The broader creative community has been vocal about its fear that generative AI will replace human illustrators, and many see any collaboration with AI startups as a betrayal of professional solidarity.
Can money change minds?
The central question posed by Pippa’s existence is whether financial compensation is enough to overcome artists’ concerns. The company hopes that as more artists see the economics of the revenue share model, they will be willing to participate. But public opinion about generative AI has grown more negative, not less, since the technology exploded onto the scene. High-profile legal cases, protest campaigns from artists, and a wave of critical news coverage have made AI the enemy in many creative communities.
Shrum and Wright are betting that they can reverse that tide. They believe their platform offers a fairer deal than the outright theft they condemn in other companies. They also think their royalty system will eventually turn early adopters into evangelists, convincing their peers to join Pippa rather than oppose it. Whether that fantasy becomes reality depends on whether artists see Pippa as a genuine partner or just another AI company trying to sanitize its practices with small payments.
The company’s focus on ethical branding is a notable departure from the industry’s usual strategy of moving fast and apologizing later. But ethics, in this context, are hard to define. Pippa may pay some artists for their contributions, but it cannot undo the damage caused by the models it builds on. And as long as the technology relies on scraped data, even the most well-intentioned revenue share will be dogged by the same old problems.
For now, Pippa remains a small player with modest subscribers at best. Its future hinges not just on acquiring more artists, but on convincing the public that generative AI can be used respectfully. If that message fails to resonate, Pippa may become another cautionary tale in the complicated intersection of creativity, technology, and compensation.
Source: The Verge News