Startup offers artists revenue share to justify AI training on their work
Pippa’s model pays illustrators per generation and a share of subscription revenue, but still relies on scraped training data.
1 source · cross-referenced
- Pippa, a text-to-video startup, markets itself as ethical by paying artists when their style is used to generate content.
- Artists receive $0.005 per image and $0.003 per second of video, plus a share of a 5% royalty pool funded by subscriptions.
- The company has 800 paying subscribers and has signed licensing/model-training agreements with four artists, with four more in talks.
- Pippa’s models are still trained on scraped internet art, though the company aims to transition to artist-provided datasets.
Pippa, a text-to-video startup, is attempting to differentiate itself in the crowded generative AI market by offering artists a revenue-sharing model. The company argues that compensating creators for the use of their work in model training can address ethical concerns that have driven legal challenges and public criticism against the industry.
Under Pippa’s model, artists are paid $0.005 per image and $0.003 per second of video generated using their style. Additionally, they receive a share of a 5% royalty pool funded by subscription revenue. The more users generate content, the larger the payments to participating artists become. Pippa’s cofounders, Hogan Shrum and Sean Wright, frame this structure as a deliberate alternative to the industry’s reliance on uncompensated data scraping.
As of August 2026, Pippa reports approximately 800 paying subscribers and has signed licensing or model-training agreements with four artists. The company is in talks to add four more artists to its roster. Participation requires artists to undergo a vetting process to confirm ownership of submitted illustrations. Approved artists gain access to a profile page to showcase their work outside the platform and may opt to use a pseudonym to avoid community backlash.
Despite its ethical positioning, Pippa acknowledges that its current models still rely on open datasets that include internet-scraped art without explicit consent. The company’s long-term goal is to transition to models trained exclusively on datasets provided by its partner artists, but this shift has not yet been fully realized. Shrum and Wright compare their approach to the music industry’s shift from Napster-era piracy to paid streaming, suggesting that ethical monetization could similarly legitimize AI-generated content.
The startup’s efforts come amid ongoing legal battles and growing resistance from artists against AI companies training on their work without permission. While Pippa’s payments to artists are small on an individual basis, the company positions them as a step toward reconciling the industry’s contentious history with creator rights.
- Aug 2, 2026 · MIT Sloan Management Review
Study finds AI financial advice improves savings and diversification but lacks nuance and rebalancing
Trust51 - Aug 2, 2026 · Simon Willison — everything
OpenAI co-founder says workers dislike AI-mediated coworker requests
Trust78 - Aug 1, 2026 · The Verge — AI
Major record labels propose excluding AI-generated songs from official charts unless 'substantially human made'
Trust75