The Era of Portrait Rights Licensing Is Coming... The Future of AI Virtual Production [AI Survival Guide]

Wonik Jang, CEO of X-on Studio
"The More You Control, the Higher the Cost... Pre-Planning Is the Solution"

The convergence of virtual production and artificial intelligence (AI) is dismantling the physical boundaries of time and space. Wonik Jang, CEO of XON Studio, has pioneered this frontier of integration. Over a 30-year career at companies such as MOFAC Studio and WYSIWYG Studio, he has led over 300 domestic and international projects. In 2020, he also established Korea's first LED-based virtual studio in Ilsan.

XON Studio Work Site / XON Studio.

He defined the current state of virtual production as a “hybrid approach that involves filming actors directly while utilizing AI.” This approach has long existed. Its origins trace back to Charlie Chaplin’s “Modern Times” glass painting technique, where virtual backgrounds were created by painting on glass, and to Stanley Kubrick’s use of front projection in “2001: A Space Odyssey” to offset shadows using reflective screens. This trajectory continued to in-camera VFX in 2020 and advanced into automated 3D model generation by 2025.

More recently, AI technologies such as Gaussian Splatting and NeRF have made it possible to remotely recreate physical locations, as seen in cases where photographs of local temples in Thailand have been reconstructed into 3D environments. Jang noted that “based on local data, this approach is minimizing the costs associated with time and space.” However, he drew a clear line, stating, “I am often asked whether this can replace all locations, but it cannot; it is valid only when certain scene-specific elements are involved.” One illustrative example was from the BTS world tour, where AI-generated seagulls were added in real-time to LED backgrounds.

The most critical challenge in merging live action with AI lies in balancing portrait rights issues and ensuring quality. Jang pointed out, “When watching content, the first thing audiences look at is the actors’ eyes,” highlighting the risk of losing emotional authenticity. He identified institutional change as the solution, predicting, “At some point, portrait rights will take the form of licenses,” with databases for famous actors’ portraits offsetting technological awkwardness.

Separate from portrait rights, AI has already brought tangible improvements in production efficiency. Amazon Prime Original’s ‘A Story of Old: Moses’ serves as a notable example. After just one week of shooting and five months of production, the unit cost was lowered by around 30 percent.

Xon Studio Work Site / Xon Studio.

Jang explained, “Production speed is three to five times faster,” and “a small team shot for a week in Los Angeles (LA) without ever having to visit the physical location.” He added, “From the outset, we designed how to utilize various technologies to deliver the final outcome,” and emphasized, “This will likely become the first major example of AI technology being leveraged at scale on an over-the-top (OTT) platform.” He also addressed the ramifications for the conservative U.S. market. “Hollywood remains highly skeptical about AI,” he said, “but this work symbolizes what kinds of changes might lie ahead for Hollywood.”

AI input models are continuously evolving. Whereas the earliest methods involved only text-based inputs, the field has progressed through multi-reference models such as Seedance 2.5 and has now reached the stage where agent AI systems—regulated by large language models (LLM) and Model Context Protocol (MCP)—control the workflow. The more sophisticated the technology becomes, the broader the range of possible controls. Jang remarked, “Increased control means higher costs, but it also ensures a certain level of output quality.”

This is where a paradox emerges. As the technology becomes more advanced, the process is increasingly reminiscent of the meticulous pre-planning found in the era of film-making. Jang predicted, “Since the cost of content generation is increasing exponentially, creators will need to hold extensive discussions in advance, much as they did during the film era, before proceeding to shoot.” He believes that only creators who understand traditional VFX rendering theories—such as camera parallax, lens angle of view, and lighting setup—will be able to achieve the highest quality results without wasting resources.

This content was produced with the assistance of AI translation services.