AI
Can artist payments convince creatives to accept AI video generation?
Pippa, a text-to-video startup, is betting that direct compensation to artists will help build trust in generative AI—though questions remain about its core technology.
The Verge
Generative AI companies have long trained their systems on artists' work without consent, sparking contentious legal battles and years of protests from the creative community. Now a new wave of startups is attempting to address these concerns by offering a different approach: paying artists when their styles generate revenue.
Pippa, a text-to-video platform, is positioning itself as the ethical alternative by building a revenue-sharing model directly into its business. Cofounders Hogan Shrum and Sean Wright argue this approach could help move past what they call the industry's problematic history and prove that AI companies can operate with artist consent rather than relying on unauthorized scraping.
How Pippa's artist compensation works#
The platform compensates creators whenever their artistic style is used to generate images or video clips. Artists receive $0.005 per image generated and $0.003 per second of video, with additional payouts from a 5 percent royalty pool drawn from the company's subscription revenue (which ranges from $14.99 to $99.99 monthly).
The vetting process requires illustrators to verify ownership of their submitted work before models trained on their art become available to subscribers. Wright acknowledged that some creators worry about community backlash, so Pippa allows artists to use pseudonyms if they choose. Currently, the platform has licensed agreements with four human artists and is in discussions with four more.
The technological compromise#
Despite its ethical positioning, Pippa hasn't fully solved the underlying problem. Its models still rely on open-source systems trained on broadly scraped internet content—meaning some unlicensed artwork remains embedded in the foundation technology. The company intends to transition toward models trained exclusively on its partner artists' work, but achieving that requires substantial capital investment most startups can't afford.
Pippa plans to integrate ByteDance's Seedance 2.5 model, which can generate up to 30 seconds of video with refinement capabilities. However, the current platform generates content largely indistinguishable from competing services—mostly generic, child-friendly animations that pale in comparison to professional studio work.
