A still image of an AI-generated clip posted by filmmaker Ruairi Robinson.
Hong Won-sik, Professor, College of Liberal Arts, ARETE, Dongduk Women’s University
Last week, a video was released that shocked the film and media industry. The video, which shows Hollywood heavyweights Tom Cruise and Brad Pitt getting into a fist fight on a rooftop, was created by ByteDance’s AI-powered video generator SeaDance 2.0. All you needed to create a fake fight video was a two-sentence prompt.
The 15-second clip not only recreated the beauty of the two stars’ respective golden eras, but also perfectly reproduced the gasps of breath and the shock of shock, shocking the entertainment and media industries. Beyond technological advancements, this video showed how AI can almost completely replace an actor’s unique presence. It was a nightmare come true in Hollywood. With predictions that traditional manufacturing methods will become extinct, it is no exaggeration to say that this video has left many people in despair.
South Korea’s Framework Law on AI went into effect in January, making it the first country to legislate regulations requiring AI-generated content to be labeled as such. Violations are subject to a fine of 30 million won (US$20,758). This can be considered a minimum safety measure to protect the public and industry from deepfakes and indiscriminate copyright law violations.
But from a media scholar’s perspective, such regulation is not the solution. They are just the first step in solving large-scale organizational problems. Just because fake content is labeled doesn’t mean the underlying copyright isn’t being violated.
Rather, the focus is not on the AI services themselves, but on the devastation they will bring to the media and entertainment industry. Signs of a crisis are already beginning to appear in the film and television industry. Last year, the number of films produced by Korean film companies decreased by more than 30% compared to previous years, and many commercial films were scrapped at the planning stage. This points to more than just an economic problem.
AI-generated content is rapidly making inroads into the short film market as well as advertising, and the entertainment industry is suffering as investment capital seeks to avoid the uncertainty of this period of technological transition.
The world is experiencing an unprecedented rate of technological development, so we can’t blame everything on governments. However, considering that South Korea is trying to develop its cultural industry into a key national industry, it is clear that it has reached a critical juncture where it cannot afford to remain silent any longer.
South Korea’s AI Action Plan, recently announced by the Presidential Council on National Artificial Intelligence Strategy, requires a thorough discussion.
The President’s Council adopted a “use first, compensate later policy” that allows AI developers to use data and content to train AI models without obtaining permission from the original creator of the data or content. The idea is to allow developers to use the data immediately without going through the associated copyright formalities that would slow down development. After the developer makes a profit, he or she will compensate the original intellectual property owner.
The advisory committee argued that global competition in AI is too intense and the country does not have time to wait for organic cooperation between AI companies and traditional content producers to develop. While this is an understandable argument, there is also an undeniable industry concern that such an exemption would effectively invalidate the core concept of copyright: the “exclusive right to grant permission.”
The council expressed support, saying it would make exceptions for industries with clear markets, such as music and news, but fundamental problems remain.
The EU’s “opt-out” approach explores a pathway for content creators to retain their intellectual property rights even after services use their data and content to train AI models. These pathways require increased accountability and transparency on the part of AI service providers and developers. However, in South Korea, there is a strong trend toward sacrificing the rights of creators in the name of nurturing the industry.
Without a basic level of transparency about how AI companies use data to train their models, promises of “we’ll make it up to you later” are likely to be empty promises. If the collection of high-quality content produced by news organizations and creators becomes subject to unauthorized “data training” conducted by AI companies, at least what remains of creativity within the media ecosystem will be depleted during this technological shift.
We cannot stand against the AI wave. But if we want to ensure that the media and content ecosystem collapses, we need macromedia industrial policy that goes beyond simple label regulation. We need to move beyond convenient concepts like “use first” and seriously consider creating minimum legal obligations for companies to be transparent about where their training data comes from.
We also need creative policies that preserve the value of high-quality content while educating the skilled workers who will be replaced by technology. We look forward to intense debate and deliberation among policymakers to create an ecosystem in which technology does not violate the dignity of content creators.
Please send your questions or comments here [english@hani.co.kr]
