How does renting a face work
In China, a new market has emerged where individuals can license their facial likeness for use in artificial intelligence-generated content, much like a photographer licenses a stock image. Platforms such as ActID and New Claw facilitate these transactions, allowing AI creators to make a person's face speak, move, and appear in various roles within short dramas or advertisements. Contributors have reported earnings ranging from $15 to $700 through these platforms, according to reports from Rest of World. For studios, these marketplaces offer virtual actors without the logistical challenges of arranging a conventional shoot. While contributors receive defined payment terms, the agreements cannot guarantee that every buyer will adhere to the intended use once the files leave the platform.
ActID reportedly has around 800 registered users, including approximately 300 individuals who are licensing their likenesses. Creators pay between 99 and 500 yuan per episode, and the platform takes a 10% commission. New Claw allows contributors to upload their own images and set a minimum licensing price of 500 yuan per image. Buyers can then use the resulting likenesses in AI projects covered by the agreement. This model differs significantly from traditional stock photography, where a single image captures one expression in one setting. A licensed AI likeness can play a recurring character, giving buyers far more creative freedom and contributors many more opportunities to encounter potential misuse.
Why does AI need so many faces
China's growing appetite for AI-generated entertainment is creating a demand for faces even when their owners never step onto a set. Figures from the China Netcasting Services Association suggest that AI was used in more than 95% of the 128,000 microdramas released during the first quarter of 2026. These microdramas are short, often low-budget productions designed for mobile viewing, and they have become a dominant form of content consumption in China. The scale of production is staggering, with studios needing to churn out dozens of episodes daily to meet audience demand. AI-generated actors provide a cost-effective solution: a single photo session can lead to an unusually long acting career, minus the trailers and uncomfortable press tours. This system allows studios to rapidly expand their character pools without the overhead of hiring multiple human actors.
The technology behind these AI likenesses relies on deep learning models that can animate static images with realistic expressions, lip movements, and body language. Companies like China's Baidu and Tencent have developed sophisticated frameworks that generate full performances from a single photograph or a short video sample. This capability reduces production time from days to hours and significantly lowers costs. However, the rise of AI-generated actors also raises questions about the future of human performers in the entertainment industry. While some actors may find new revenue streams through licensing their face, others may face reduced opportunities for traditional roles. The demand for faces is not limited to dramas; it extends to advertising, virtual influencers, and even customer service avatars, creating a broad market for facial assets.
Can you get your face back
Chinese law requires explicit consent before biometric data can be processed and prohibits fabricating someone's likeness without permission. Face-licensing platforms add contracts and technical restrictions intended to keep buyers within the agreed uses. However, enforcement is significantly messier than the legal framework suggests. The Guangzhou Internet Court has handled around 700 AI-related disputes over three years, including cases involving face-swapping and copyright infringement. These cases highlight the challenges of policing digital content once it leaves a controlled environment. For instance, a licensed avatar might be used in a pornographic video or a political advertisement without the contributor's consent. Even with contractual protections, tracking unauthorized use across the vast expanse of the internet is difficult.
Technical measures such as digital watermarks and blockchain tracking have been proposed, but they are not foolproof. Watermarks can be removed, and blockchain records only ensure transaction history, not subsequent misuse. The platforms themselves may lack the resources to actively monitor all uses of their licensed assets. This situation creates a gap between legal rights and practical enforcement. Anyone tempted to license their face to AI should look beyond the initial fee and investigate how each platform monitors unauthorized use. A contract defines who has permission to use a likeness, but it cannot make every rogue version disappear. The burden often falls on the individual to prove infringement, which can be expensive and time-consuming.
The rise of AI-generated microdramas
The explosive growth of microdramas in China is a key driver of the AI face licensing market. Microdramas are short-form serialized stories, often with dozens of episodes that each last just a few minutes. They are designed for vertical mobile viewing and are typically monetized through in-app purchases or advertisements. The production cycle for these dramas is extremely fast; some studios release multiple episodes per day. Traditional filming methods cannot keep pace with this output, so AI-generated content has become a necessity. By using licensed faces and automated animation tools, producers can create entire episodes without any physical actors, sets, or crew. This approach has attracted significant investment, with some AI drama production companies raising millions in venture capital.
The Chinese government has shown interest in regulating this space. In 2025, the Cyberspace Administration of China released guidelines requiring clear disclosure when AI is used to generate or manipulate content. These regulations aim to prevent deepfake fraud and protect consumers from deceptive practices. However, they also create compliance burdens for platforms and studios. The face-licensing platforms themselves operate in a gray area: they provide tools for legitimate uses but cannot fully prevent misuse. The Guangzhou Internet Court's caseload indicates that disputes are becoming more common as the technology spreads. Legal experts suggest that clearer liability rules for AI-generated content are needed to protect both contributors and consumers.
Technical and ethical challenges
From a technical standpoint, creating convincing AI actors requires high-quality training data and sophisticated neural networks. Most platforms use a generative adversarial network (GAN) or diffusion model to produce realistic video from a single image. The results can be eerily convincing, especially for static poses or simple expressions. However, generating consistent performances across multiple scenes and emotions remains challenging. Some platforms offer templates for different character types—such as a scholar, a warrior, or a doctor—that can be applied to any licensed face. This modular approach allows buyers to mix and match faces with personalities, but it can also lead to ethical concerns. For example, a person's face could be used to portray a villain or a victim in ways they find objectionable.
Ethically, the practice of licensing one's face raises questions about identity, consent, and exploitation. While contributors receive compensation, the asymmetry of power between individual licensors and large studios can lead to unfair contracts. Some platforms include clauses that allow perpetual use without additional payment, leaving contributors locked into agreements they cannot easily exit. Moreover, the lack of standardized industry practices means that contributors may not fully understand how their likeness will be used. Advocacy groups have called for transparency rules that require platforms to disclose all intended uses upfront. There is also the issue of deepfake fraud, where malicious actors use licensed faces to impersonate individuals for scams or disinformation. The same technology that powers AI dramas can be repurposed for harmful purposes, making it imperative for platforms to implement robust security measures.
Comparison with stock photography
The face-licensing model borrows heavily from the stock photography industry, where photographers earn royalties from repeated use of their images. However, the stakes are higher for faces because a likeness is more personal and cannot be easily changed. In stock photography, a model may be photographed once and the image can be sold many times, but the model retains control over their appearance in other contexts. With AI licensing, the same face can be animated in infinite ways, effectively creating a digital twin that operates independently of the real person. This amplifies both the opportunity for revenue and the risk of misuse. Some contributors have reported that their licensed face appeared in content they found offensive, such as horror dramas or political propaganda. Legal recourse is possible but slow, especially if the unauthorized use occurs outside China.
Platforms are beginning to address these concerns by offering more granular control. For instance, ActID allows contributors to set limits on the type of content their face can be used in, such as excluding adult or political themes. New Claw provides a review process where contributors can approve or reject each use case. However, these features rely on the platform's ability to enforce them, which is limited by the decentralized nature of the internet. As the market grows, we may see the emergence of industry standards similar to those in stock photography, such as exclusive vs. non-exclusive licenses and time-limited agreements. The key difference is that AI likenesses can be updated and improved over time, meaning that a license granted today could continue to generate new content for years.
Global perspective on AI likeness licensing
While China is the current epicenter of AI-generated microdramas, similar practices are emerging in other parts of the world. In the United States, companies like Metaphysic and Rosebud AI offer services that allow individuals to create digital avatars for use in videos and games. However, the regulatory environment is fragmented. The European Union's AI Act classifies deepfakes as high-risk and requires transparency labeling, while California has laws governing the use of a person's name, image, and likeness. In China, the combination of a huge domestic market, relatively lax enforcement, and a cultural acceptance of AI has accelerated adoption. The country's microdrama industry produced over a million episodes in 2025, and AI-generated content is expected to dominate by 2027. This scale creates a feedback loop: more demand leads to more faces being licensed, which leads to more advanced AI models, which further drives down costs and expands applications.
For individuals outside China, the possibility of licensing their face for global AI content is becoming a reality. Some platforms accept international contributors, though payment and legal protections vary. The lack of international harmonization in biometric data laws poses risks for cross-border licensing. A person in Europe might license their face to a Chinese platform only to find that their data is not protected under the GDPR. Experts recommend that contributors consult with legal advisors before entering into any agreement with foreign platforms. As the technology evolves, the line between human and AI actors will continue to blur, raising profound questions about the nature of performance, identity, and ownership. The platforms that succeed will be those that balance innovation with responsible stewardship of the human faces they digitize.
Source: Digital Trends News