Cineverse, the tech-forward indie studio behind the Terrifier franchise, has launched Matchpoint Reel Visuals AI, a new artificial intelligence rights management service to connect content owners interested in non-exclusive revenue sharing to feed AI training models.

The service aims to help owners of libraries, feature films, television shows, podcasts and more take advantage of growing demand from AI companies to acquire audio and video content to train artificial intelligence models at scale.

Matchpoint will work with content creators and rights holders from independent production companies, studios, content aggregators and AI training rights agents to connect with the biggest AI platforms and get paid to contribute their video to AI training, which informs how effective a generative AI platform can be.

Matchpoint also eliminates the associated delivery costs that often make the opportunity to profit in this space unviable. “We are a one-stop solution,” Cineverse CEO Chris McGurk tells Deadline.

Cineverse has been using Matchpoint to manage its own content and is now expanding to license it to third parties. It’s put deals in place over the last six to eight weeks with content producers for about 350,000 hours of video and audio fare. “We saw the market opportunity and realized that we are well positioned,” he said.

Cineverse, with multiple channels and 70,000 library titles, podcasts and more, has been a content aggregator for some time so this seemed to be “an absolutely natural” next step.

The key, he said, are guardrails. These deals are  just for AI training. “The large language models can’t use it for anything other than AI training.” No “web scraping” allowed, no using the content in any other way.

AI is highly controversial in Hollywood as elsewhere for the threat it poses to livelihoods and, some fear, the creative industry as a whole. Authors and others have sued Microsoft-backed OpenAI, Meta’s Llama, Jeff Bezos’ Perplexity and others for using copyrighted material for training as well for spitting it out directly with or without permission or attribution.

Last week, the WGA slammed Hollywood studios for not protecting copyrighted works used in generative AI training models, urging them to “come off the sidelines.”

McGurk declined to discuss details of the revenue sharing pacts, which are different for every deal and depend on what the AI model is looking for on any given day, which could, for instance, be lots of Spanish-language films, or 2,000 hours of horses running in a field, McGurk said. According to stats from a recent report the company cited, the global AI training dataset market was valued at $2.6 billion in 2024 and is projected to reach $10.3 billion by 2030.

Cineverse COO and CTO Tony Huidor described the company as “stepping into an important leadership role by establishing a much-needed framework that will ensure that intellectual property rights are respected and content owners are properly compensated.”

“It is imperative that we not only work to resolve emerging market issues but also advocate for ethical AI content procurement practices by pursuing equitable compensation for the use of video in AI training,” he said.