Is paying artists enough to convince them to embrace AI?

For years, illustrators have sounded the alarm about generative AI startups training models on their work without permission, labeling the practice as theft. In response, a new wave of startups like Pippa is emerging, marketing services as more ethical by promising to pay artists directly.

Pippa's Revenue Share Model

Pippa’s main product is text-to-video generation, but its business model is designed to compensate artists. Every time a subscriber generates content in a style derived from a human artist’s work, that artist receives a payment. Founders Hogan Shrum and Sean Wright believe this focus on ethics will attract customers and set a new industry standard.

The individual payments, however, are modest. Artists receive $0.005 per image and $0.003 per second of video. Additionally, they share a 5 percent royalty pool funded by subscription revenue, which ranges from $14.99 to $99.99 monthly. Pippa compares its structure to Spotify—a choice that may raise eyebrows given the long-standing complaints from musicians regarding the streamer's payout rates.

The Data Dilemma and Artist Apprehension

Despite its goals, Pippa’s technology still relies on open models that were initially trained on datasets scraped from the internet without creator consent. While the company aims to move toward purely in-house datasets, it currently cannot fully solve the issue of creative theft inherent to generative AI. Wright acknowledges this, referring to the industry's "bloody history" of unauthorized training.

Convincing artists to join remains an uphill battle. Pippa currently has only four signed artists, with four more in talks. Many creators fear being ostracized by their peers for collaborating with AI firms. To mitigate this, Pippa allows artists to submit work under pseudonyms to avoid being seen as "crossing the picket line."

Visual Quality and Future Tech

Currently, the platform's output lacks distinct artistic flair. The site is largely filled with kid-oriented content attempting to mimic Pixar’s aesthetics, rather than showcasing unique independent styles. To improve, Pippa plans to integrate ByteDance’s Seedance 2.5 model, which offers better fine-tuning capabilities. However, its long-term viability depends on whether it can successfully pitch its economic model to a skeptical creative community.