Clint Stinchcomb
President and CEO at CuriosityStream
Over the next five years, AI model development, model refresh cycles, geographic expansion, enterprise fine-tuning, education applications, agentic systems, and multimodal search should all support continued appetite for premium licensed corpora. For AI license partners, as their model sophistication grows, so does the need for more video inputs. Developers require large volumes of high-integrity, rights-aware training inputs. Premium broadcast video, clean audio, scripts, captions, study guides, metadata, and derivative assets have utility well beyond entertainment viewing. They help train, tune, evaluate, ground, and improve multimodal systems. The more advanced models become, the more they need high-quality, structured, legally licensable data rather than undifferentiated scraped material. Key to note that rights-cleared, structured content will become more valuable over time, not less. There's plenty of media on the open web, but much of it is noisy, duplicative, poorly labeled, low quality, or legally ambiguous.