AI digital people pretend to be doctors to sell drugs, and supervision needs to penetrate the chain of responsibility

Tang Chuanyan

According to a CCTV news report on September 28, it is very common for AI digital people to "endorse" advertisements on online platforms. Reporters browsed multiple online platforms and found that a large number of AI digital people who appeared as "medical workers" and "field experts" "endorsed" to promote related drugs and health products such as eye diseases and gray hair care. Many promoters claimed in the video that the products were "effective in personal testing." Such videos generally have the problem of exaggerated propaganda, and some videos also show so-called "scientific experiment" scenes.

More and more digital people are imitating doctors, which is directly related to factors such as the sudden drop in the cost of fraud. In the past, when someone was asked to play the role of a doctor to shoot a sales video, one had to prepare the venue, props, and speaking skills; now, by inputting a few words, the scene can be generated with one click. The white coat, badge, and laboratory background are all matched by an algorithm, and the age and professional title can also be set at will. Some software sets no threshold for the image of medical staff, supports batch production, precise delivery, and even provides the function of removing watermarks, making it more difficult for patients to identify. The cost is only a few yuan, but the circulation can reach hundreds of thousands. This contrast drives counterfeiters to continue to enter the game.

An AI digital person has neither the qualifications to practice medicine nor the ability to perceive, but it is ridiculous to talk about curative effects and personal testing as a doctor. The Advertising Law clearly stipulates that advertisements for medical treatments, drugs, and medical devices shall not use advertising spokespersons for recommendations or certifications, and the images of medical personnel shall not be used in commercial advertisements. Even real doctors are subject to strict restrictions. Digital humans do not have the conditions for real trials, let alone talk about efficacy or endorse medical products. The so-called "personal testing" is essentially anthropomorphic false propaganda.

The biggest victims of the above-mentioned AI digital human advertisements are of course patients. If patients believe in marketing rhetoric and give up regular treatment to buy so-called miracle drugs, they may spend a lot of money and delay their condition. Digital people can broadcast live 24 hours a day, and their online time and reach exceed that of real-person propaganda. Groups with weak health literacy often become the targets of harvest without knowing it.

The "Measures for the Supervision and Administration of Live Streaming E-commerce" implemented in February this year clarified that those who use artificial intelligence-generated character images and videos to engage in live broadcast e-commerce activities should continue to be prompted for AI generation. However, it must be clear that even if there are relevant reminders, it does not mean that all parties in the chain can stay out of the matter. Once something goes wrong with a product, since digital people cannot take responsibility, the responsibility must pass through to the real person behind the screen. Advertising is mainly responsible for the authenticity of the content, and the producer and publishing platform are responsible for auditing. Who writes the script, who operates the account, who delivers it to patients, and who ultimately makes a profit, every link must be traceable. AI-generated software is also obliged to regulate usage behavior and intercept content related to drug efficacy and other aspects in advance. Governance cannot only rely on removal from the shelves after the fact. Only when relevant platforms upgrade screening methods and regulatory authorities take heavy action can counterfeiters feel that it is not cost-effective and dare not commit the crime again.

In fact, AI digital humans are not useless in the medical and health field. Popularizing health knowledge, explaining how to use it, and answering common questions are their strengths. We hope that the medical and health industry will use new technologies correctly and well to further break down information barriers and reduce the burden on patients. We also hope that all parties will work together to solve the problem and draw red lines and clarify the boundaries for digital human advertising.