Pippa, a text-to-video AI startup, is launching with a promise to pay artists royalties for training data instead of scraping it without permission. The move comes as illustrators wage legal war against giants like Google and Meta over unauthorized use of their work. Artists recently launched coordinated legal action against Google, Meta, and Anthropic, seeking damages for what they characterize as systematic copyright infringement at massive scale. The cases could take years to resolve and potentially billions in damages are at stake. Industry defenders have countered that such training constitutes fair use and that restricting it would cripple innovation. According to arguments made by OpenAI to the Trump administration, limiting training data access could cause the U.S. to lose the AI race to China. Pippa claims to have built its dataset through licensed partnerships with artists who receive ongoing royalties. The model resembles how Spotify or Apple Music compensate musicians, though the economics of AI training create different dynamics. Where streaming services pay per play, AI training involves a one-time ingestion of creative work that then influences countless generations. How exactly Pippa structures those payments—whether as upfront licensing fees, per-generation micropayments, or percentage-based royalties—remains unclear from public statements. Pippa joins a growing wave of startups marketing themselves as the ethical alternative to Big Tech's approach, wagering that creators and brands will pay premium prices for AI tools that compensate artists.