ÉTAT HER

Pippa Pitches Illustrators on a Spotify-Style AI Royalty Deal — Only 4 Artists Have Signed On

Generative AI startup Pippa wants to replace unauthorized data scraping with royalty sharing. Subscriptions run up to $99.99/month, but artists are paid just $0.005 per image — and the underlying model is still built on scraped web data.

Pippa Pitches Illustrators on a Spotify-Style AI Royalty Deal — Only 4 Artists Have Signed On

Pippa's two co-founders, Hogan Shrum and Sean Wright, liken their company to the music industry after Napster — piracy got shut down, and Apple pulled musicians back onto the paid track with 99-cent songs. "We want to do the same thing for AI video generation," they say. It's an inspiring analogy, but they picked the wrong reference point: Spotify's royalty system is exactly the model musicians have been complaining about for years. And yet Pippa's own press materials directly benchmark its payout structure against Spotify.

Pippa launched this past May and currently has about 800 paying subscribers, with monthly plans ranging from $14.99 to $99.99. But the number of human artists the company has actually signed and secured training consent from? Just 4, with 4 more in talks. Illustrators who want to join must first pass a multi-step verification process to prove the submitted work is genuinely their own. Once approved, a model trained on their style becomes available to subscribers, and the platform gives the artist a personal profile page to redirect traffic back to their own work. Recognizing that many illustrators fear being seen by peers as "crossing the picket line," Pippa also lets artists participate under a pseudonym.

The real issue is compensation itself. Artists earn $0.005 per generated image and $0.003 per second of generated video. On top of that, 5% of subscription revenue goes into a royalty pool shared among signed artists — the very mechanism Wright describes as being "just like Spotify." The gap between what users pay and what artists actually receive is precisely the long-standing complaint leveled at Spotify — and Pippa is voluntarily inviting the comparison.

More fundamentally, the models Pippa currently uses to generate content are still trained on "broader web data," meaning the foundation remains built on scraped material obtained without consent. The company says it plans to gradually replace training data with content contributed by its own signed artists. By contrast, Ben Affleck's AI company InterPositive chose to build a fully proprietary, original dataset from scratch — but that requires massive upfront capital that most startups can't afford, and the resulting tools tend to serve narrow use cases rather than general consumers.

Browsing the actual content generated on Pippa's platform, it's hard to spot much difference from other text-to-video services: mostly children's-style animation that imitates a Pixar look without quite nailing it. It's difficult to identify any video that genuinely carries the distinct touch of an independent illustrator. Pippa is next planning to integrate ByteDance's Seedance 2.5 model, which can generate video clips up to 30 seconds long and allows fine-tuning without full regeneration, as a hook for subscribers — but that technology is available on other platforms too. For Pippa to truly differentiate itself, it still needs to sign on more local artists.

Shrum and Wright believe they can convince more illustrators to join, and that their current roster of signed artists will help spread the word to peers. But for any of this to work, the illustration community's perception of generative AI needs to shift first.

Related

New Anthropic Policy Bars "Prolonged, Gratuitous Abuse" of AI Models, With Election Rules Also Tightened
Living

New Anthropic Policy Bars "Prolonged, Gratuitous Abuse" of AI Models, With Election Rules Also Tightened

In its annual usage policy update, Anthropic introduced a clause prohibiting users from inflicting "prolonged and gratuitous" abuse on Claude and other models, while election-related rules were also tightened—sparking renewed debate over whether AI should be regarded as a sentient subject.

A "Paper-Like" Screen Built with Corning Etching Tech Debuts on Amazon's 12-Inch Alexa+ Pro Tablet
Living

A "Paper-Like" Screen Built with Corning Etching Tech Debuts on Amazon's 12-Inch Alexa+ Pro Tablet

The flagship 12 Pro showcases a Nanomatte etched display co-developed with Corning, as Amazon rolls out three new Alexa+ tablets—12-inch, 11-inch, and 8-inch models. Pricing starts at $230, with availability in the US, Canada, and Mexico beginning October 14.

Microsoft Puts Agentic AI at the Heart of Windows Search as Jensen Huang Looks On and Davuluri Takes the Stage
Living

Microsoft Puts Agentic AI at the Heart of Windows Search as Jensen Huang Looks On and Davuluri Takes the Stage

At a San Francisco event on October 7, Satya Nadella and Pavan Davuluri unveiled Microsoft's Windows and Surface plans while Jensen Huang watched from the audience. The centerpiece is a deeper integration of Copilot into the search bar and local files, set to arrive this fall.

Ring's Fix for Smart Lock Skeptics: A Manual Crank When Power Runs Out
Living

Ring's Fix for Smart Lock Skeptics: A Manual Crank When Power Runs Out

According to Ring founder Jamie Siminoff, fear of dead batteries has long held smart locks back. The company's answer is a hand-crank generator built into its upcoming Smart Lock, letting users turn a handle for emergency entry. Launch is set for early 2027 at $249.

OpenAI's 722 Math Papers: Single-Prompt Breakthroughs or Unverifiable Claims?
Living

OpenAI's 722 Math Papers: Single-Prompt Breakthroughs or Unverifiable Claims?

OpenAI published 722 manuscripts on GitHub citing progress on 372 difficult math problems, most generated by feeding one prompt to one AI agent. The underlying model remains undisclosed, and neither the prompts nor computation times were shared, raising doubts about verification.

Your Google One Tier Now Decides Access: Google Launches Experimental Game-Building Platform Playground, Just Type to Create
Living

Your Google One Tier Now Decides Access: Google Launches Experimental Game-Building Platform Playground, Just Type to Create

Access to Google's new AI game creation tool, Playground, is tiered according to your Google One membership level, and for now limited to US users aged 18 and older. The platform lets anyone generate 2D or 3D games from a few text prompts alone, without coding or drawing skills.