AI ‘doctors’ are spreading fake health tips on YouTube. Regulators can’t do much to stop them.

AI-generated videos featuring fake doctors dispensing health advice are proliferating on YouTube, and there's very little the authorities can do about it.

Current laws provide limited grounds to regulate such content unless it directly advertises products, and content creators are exploiting that loophole. On a channel with more than 36,000 subscribers, videos claiming to be affiliated with a well-known university hospital are churned out with seemingly useful information related to health issues affecting older adults. 

"You're probably wondering which health videos on YouTube you can trust," says an AI-generated doctor in one of the videos by the channel. "Everything covered here has been verified by a university hospital."

Some of the videos include "Seven Reasons an Internal Medicine Doctor With 25 Years of Experience Drinks Vinegar Every Day" and "A 67-Year-Old Avoided Dialysis in Three Months with Honey" — all seemingly harmless, and all so attractive to health-conscious adults.

However, the university hospital that the videos claim to be affiliated with said it has no connection to the channel.

"The doctor appearing in the videos is not a real physician but an AI-generated virtual character," a hospital spokesperson said.

Similar AI-generated health content featuring fake medical professionals has spread across online platforms, but regulators have struggled to respond because laws currently on the books primarily target conventional advertising.

Data submitted by the Ministry of Food and Drug Safety to Rep. Lee Ju-young of the Reform Party showed there were no cases of false advertising by channels dispensing medical advice with AI between January and May.

The government introduced an AI-generated content-labeling system and punitive damages for false or exaggerated AI-generated advertising in January, saying the measures would curb deceptive advertising. However, not a single violation has been detected.

"We had not identified any advertisements using AI doctors," the ministry said.

Health videos featuring AI-generated experts have become widespread on YouTube.

One channel with more than 74,000 subscribers primarily uploads eye-catching videos with titles such as "A Urologist's Warning About This Bedtime Habit" and "Three Dangerous Fruits That Damage Your Kidneys After Age 60" featuring what appears to be an AI-generated physician. The channel also posts affiliate advertising links in its video descriptions and earns commission fees.

Current law makes it difficult to regulate such content. The Ministry of Food and Drug Safety's enforcement authority under the Act on Labeling and Advertising Foods is limited to advertisements for food and health functional foods. If a video does not directly sell food or health supplements, it generally falls outside the scope of the ministry's enforcement authority. The Medical Service Act also does not specifically prohibit using AI-generated doctors to attract viewers with false or misleading health information.

Instead, many of these videos generate revenue through YouTube advertising, view counts and affiliate marketing, effectively turning misleading health information into a business.

"We plan to inspect AI-generated video advertisements featuring fake doctors and other AI-generated figures later this month," said the Health Ministry.

"Even if a video is not a direct sales advertisement, an operator may still be held responsible for unfair labeling or advertising if it is determined to have sales intent, such as linking viewers from YouTube to a product sales page."

The regulatory gap has contributed to the rapid spread of AI-generated health content featuring impersonations of doctors and other experts.

An analysis by the Korea Health Promotion Institute of the 100 most-viewed YouTube videos related to older adults' health posted during April found that 42 videos used AI-generated content. Among them, 24 impersonated doctors or other experts.

"Many of the videos were found to exaggerate health information or present claims lacking scientific evidence," the institute said.

Experts warned that such videos could pose direct risks to public health by spreading inaccurate medical information.

"There is no shortage of dangerous content among AI doctor videos. The government and civic groups need to establish a joint monitoring system," said Oh Sang-woo, a professor of family medicine at Dongguk University Ilsan Hospital.

"Consumers are highly likely to mistake AI doctors for real experts, and the information they provide can lead to actual harm," said Jeong Ji-yeon, secretary general of the Consumers Union of Korea. "As the regulatory vacuum continues, the government and the medical community need to respond more actively."

"The government introduced an AI-generated content labeling system and punitive damages, but the absence of a single case over the past five months shows the system is not functioning properly," the Reform Party's Rep. Lee said. "Comprehensive measures should be pursued by revising the Medical Service Act, the Act on Labeling and Advertising Foods and the Act on Promotion of Information and Communications Network Utilization and Information Protection together."

BY CHAE HYE-SEON [[email protected]]