An explosive rise in AI-generated short dramas has spawned a booming new industry centered on facial authorization, with viral topics like "ordinary people selling their face for 500 US dollars in three days" and "celebrities cashing in on their likeness" capturing widespread attention. Alongside this trend, concerns over the abuse of portrait rights, leaks of personal information, and unchecked authorization have come to the forefront. Many users are left wondering exactly where their facial data goes, how it will be used, and how they can put a stop to it. Legal experts were consulted to break down the legal risks of these facial licensing deals.

This "face-selling" craze involves everyday individuals and actors granting one-time commercial rights to their likeness for use by producers of AI short films and digital advertisements. On various platforms, after completing identity verification and a facial scan while agreeing to sensitive information consent forms, a user can get their "face" listed for trade within just 30 minutes. These platforms allow people to choose their own acting genres and set their own prices, with ordinary users earning about 100 yuan per short video authorization, though some platforms enforce a minimum quote of 500 yuan. Professional actors can command between 500 and several thousand yuan, with some reaching as high as 6,000 yuan per gig. After a transaction, the platform and the rights holder split the earnings, with one platform reporting over a hundred matched orders totaling nearly 70,000 yuan in just two months of operation.

Lawyers explain that the market's explosive growth is fueled by the AI short drama sector, which has already surpassed 22 billion yuan in scale during the first five months of this year, with AI micro-dramas representing over 95% of that output. Since native AI-generated faces often look robotic, real human digital avatars offer a more authentic feel, driving up demand for facial authorizations. However, the industry is already seeing signs of oversupply and dropping prices. More critically, since a face is considered sensitive biometric data, some platforms are collecting excessive information, hiding risks of data breaches and unauthorized model training.

One legal expert calls this a "new industry chaos where technology has outpaced regulation," noting that platforms attract users with small payments, simplifying the authorization process to mass-collect facial data. This data is then used to generate AI digital humans and short drama characters at low cost, shifting the burden of risk onto the individual.

The two lawyers hold differing views on whether they would personally sell their own face. One firmly states he would never participate due to the unacceptable risk-to-reward ratio, explaining that people receive a paltry sum upfront while shouldering potential damage to their portrait, privacy, and reputation. He warns that most users cannot fully comprehend the complex terms and cannot foresee how their data will be used or transferred later, creating costs that no amount of money can offset. The other lawyer believes that, if platforms strictly follow the existing legal framework for protecting sensitive personal information and honor users' rights to withdraw consent, he would consider it. He sees the nascent market as having a potentially broad future, allowing people to unlock their personal image value and expand their career opportunities.

Both lawyers agree that the current foundation for these facial authorization deals is weak, hiding multiple legal risks, with damage that can be continuous and irreversible. The most prominent risk is loss of control over the authorization itself. The law requires that processing sensitive personal information like faces requires separate, explicit consent that specifies the purpose, scope, and duration, and prohibits blanket authorization. Yet many platform agreements include clauses like "perpetual use" and "transferable to third parties." A user may intend to license their face for only one short drama, but their data could be circulated endlessly for ads, live streams, or even gray-area activities. Even if consent is later withdrawn, it is unclear whether AI models already trained on that face or the derivative videos generated from it can be completely purged.

A more alarming scenario is when facial data leaks into black-market channels and is used to create telecom fraud videos. Victims could then be dragged into criminal investigations, wasting significant time and energy trying to prove their innocence. Beyond contract traps, the risk of data breaches is severe. Some platform agreements use broad terms like "perpetual authorization" and "all AIGC uses" while being vague on crucial points like revocation rights, model deletion, and breach compensation. Since a face is unique, once the biometric data stored on a server is exposed, the damage from leaked facial features is irreversible.

Finally, the burden of legal action falls heavily on the individual. While platforms violating rules face administrative penalties and potentially criminal charges for infringing on citizens' personal information, the reality is that many face-trading activities are hidden within small apps or private QQ and WeChat groups. When an ordinary person's rights are violated, they often cannot even identify the responsible party or gather valid evidence, making the cost of pursuing legal recourse extremely high.

To help ordinary people avoid these pitfalls, the lawyers offer professional advice centered on a principle of "avoiding the source." The primary recommendation is to avoid commercial facial licensing altogether, especially any deals brokered privately in WeChat groups, which have no compliance guarantees and should be flatly refused. For those who insist on licensing their face, it is critical to strictly review the contract before signing, specifying that the use is limited to a particular type of AI short drama and explicitly excluding any application in advertising, product promotion, or false publicity. The agreement should reject any perpetual authorization clause, setting a limited term of one to two years, requiring renewed consent for any extension, and explicitly forbidding sub-licensing to third parties. It should also include a written unilateral withdrawal right, detailing that the platform must delist the digital avatar, delete the original facial data, destroy training models, and outline the handling of existing works and penalty clauses. The authorization must be a standalone consent, not buried inside a general user agreement.

Before signing, users must verify the platform's registered business qualifications and ensure all verbal promises are committed to writing. In addressing the common complaint of "easy to authorize, hard to cancel," legal experts confirm that under the Personal Information Protection Law, citizens have a statutory right to withdraw consent and request data deletion. Platforms are legally obligated to comply and cannot reject the request based on their own contract terms. A practical step-by-step action plan was provided for users: first, save screenshots of the authorization page, the agreement, and transaction records to confirm the platform's operating entity. Next, send a written request to the platform's official email demanding immediate termination of all commercial authorizations, deletion of original facial data and derived AI models, stopping any third-party sub-licensing, and requesting written confirmation of the actions taken. If the platform refuses, users can escalate the issue by filing complaints through cybercrime reporting channels like 12377, the consumer rights hotline 12315, or the local cyberspace administration bureau to leverage regulatory pressure.

Beyond these actions, daily vigilance is essential: refuse any non-essential facial collection requests and never tick boxes that grant perpetual or all-encompassing authorization. Facial licensing is a new transaction model born from the AI content industry, unlocking portrait assets but creating a growing conflict between personality rights and biometric data protection. While current judicial opinions have begun to regulate AI face-swapping and misuse of personal portraits, and new draft regulations for digital virtual human services require separate consent for face modeling, issues like crude industry contracts, difficulty in monitoring infringement, and problems with deregistration persist.

There is a call for regulators, judicial departments, and platforms to work together to refine the rules for facial authorization, withdrawal, and cancellation to reduce violations of personality rights. The market cannot simply chase commercial monetization. All three parties the platform, the purchaser, and the rights holder must respect legal boundaries. Finding the balance between the innovative growth of the AI content industry and the protection of personal biometric information remains a crucial and ongoing challenge in the AI era.