As generative AI spreads, content companies around the world are resetting their standards for how AI-generated material can be used. Content distribution platforms such as YouTube and Meta have largely focused on labeling or disclosing AI-generated or AI-altered content. Netflix, by contrast, has gone further. It has established separate guidelines requiring prior disclosure and approval when generative AI is used during the production process. The move is seen as a way to manage copyright and portrait-right risks in advance, given that Netflix directly owns the intellectual property rights to its original content.
/ Gemini-generated image
According to the Korea Information Society Development Institute (KISDI), Netflix distributed its “Using Generative AI in Content Production” guidelines to production partners and vendors in August last year. The document says that if generative AI is to be used in content production, the plan must be shared in advance with the relevant Netflix personnel and used only within the principles set out by the company.
This goes a step beyond the generative AI policies adopted by major domestic and overseas content platforms including Google YouTube, Meta, Naver, Kakao and SOOP. Most content platforms have focused on requiring disclosures indicating whether posted content was created using AI.
YouTube requires creators to disclose when they upload AI-generated content that appears realistic. Meta also applies “AI Info” labels on Facebook, Instagram and Threads. Korean platforms such as Naver and Kakao have taken a similar approach. Under South Korea’s Framework Act on Artificial Intelligence, which took effect in January this year, content generated by AI must be labeled as such.
Netflix applies especially strict standards when AI-generated material is included in final content. If generative AI is used in production for matters involving third-party copyrighted works, unpublished materials, or an actor’s likeness or voice, prior consultation with Netflix and written approval are required. Separate principles covering copyright protection and performer rights must also be followed.
Experts say Netflix issued such guidelines because its business model differs from content platforms such as YouTube, Instagram, SOOP, Chzzk, Clip, Naver Webtoon and KakaoPage. Platforms like YouTube mainly help users upload, post and distribute content on their own. As a result, they focus on preventing incidents involving AI-generated deepfake videos and similar risks.
Netflix, on the other hand, directly owns the copyrights to its original content. That means if AI-generated output is included in the final version of a work and results in copyright or portrait-right infringement, liability could fall on Netflix. In effect, the guidelines amount to a form of quality control.
Netflix has already faced real controversy over AI-generated output. Its 2024 original documentary What Jennifer Did included images suspected of having been generated by AI.
At the time, foreign media criticized the documentary’s authenticity and quality, citing images such as a woman whose fingers appeared distorted. Netflix later created and distributed the guidelines. Earlier, during its earnings announcement in July last year, Netflix said it had used AI for visual effects in its original drama The Eternaut.
Han Jeong-hoon, head of K-EnterTechHub, said, “Netflix’s emphasis on prior discussion and approval when using AI can be seen as an effort to control the quality and credibility of original IP released under its own brand, going beyond simple watermarking. By contrast, platforms such as YouTube and Instagram focus less on the substance and quality of individual IP and more on whether AI content across the platform is clearly identified and properly disclosed so that users are not confused into thinking it is real.”
기자의 전체기사
저작권자 © IT조선
키워드
#Netflix
#AI
#Guidelines
댓글
0