Singaporean director and AI consultant Ehuanglu demonstrates the potential of using generative AI to create hyper-realistic digital actors, whose visual expressiveness and human-like resemblance are questioning the future of traditional professions in film and dubbing.
What Happened
Director Ehuanglu (el.cine) published works on social media created using generative video and audio tools such as Sora, Kling, Runway, and ElevenLabs. These demonstrations show the ability of AI to create digital twins with a high level of emotional expressiveness and visual similarity to real people.
Context
The technology stack, including video generation models (Sora, Kling, Runway) and audio synthesis tools (ElevenLabs), is forming a new content production ecosystem. However, a technological gap is observed: while the visual component is achieving hyper-realism, the quality of automatic dubbing—especially in non-English languages, including Russian—still faces serious problems with emotional synchronization and artifacts.
Why It Matters for the Industry
The rapid development of generative content creates direct competition for live actors and dubbing studios. This is forcing the industry to reconsider the legal and ethical aspects of using digital twins, as well as to seek new models for protecting intellectual property and standards for evaluating the multimodal consistency of AI content.
Why It Matters for Users
The boundary between real and synthetic video content is becoming virtually indistinguishable. For ordinary users, this means the need to develop media literacy skills and critical perception of information to distinguish generated images from real recordings.
What Is Not Yet Known / Limitations
There is a quality gap between visual generation and audio synthesis for non-English languages, which limits the possibilities for immediate global production automation.
Sources
Author
Look at AI, Editorial Team