Don't Forward Again! BBC Investigates the Truth Behind 'AI Fake Doctors': Most Popular Among Taiwan's Elderly, YouTube Becomes a Breeding Ground for Misinformation

In a classroom at a senior day care center in Taipei, elderly participants averaging nearly 80 years old are focused on their smartphone screens. They watch a man identifying himself as 'Dr. Zheng,' wearing a white coat and stethoscope, calmly discussing 'sarcopenia in the elderly' and claiming that consuming protein after age 60 is ineffective. This video is familiar to He Yiqing, one of the attendees. He had previously received it from a friend and casually forwarded it. But today, he sees it differently. The AI assistants Gemini and ChatGPT, which he recently downloaded, told him that 'Dr. Zheng' does not exist—his image was AI-generated, and the content is a mix of real and fabricated health information pieced together by AI. 'I send many videos to group chats every day. Only now do I realize some aren't real,' He told BBC Chinese.

The Rise of 'AI Doctors'

Eighty-one-year-old He was attending a smartphone app course designed for seniors. The instructor taught them to use AI assistants like Gemini and ChatGPT to plan travel itineraries, analyze health check reports, and upload commonly viewed AI videos to verify their authenticity. Li Ziqiang, the community college teacher leading the class, told BBC Chinese that over the past year, AI-generated videos have spread widely among elderly communities and messaging app groups in Taiwan, raising alarm. The most common format features a virtual 'doctor' or 'expert' delivering information to viewers. 'The content sounds important and relevant to their daily lives, but upon closer listening, you start to question: Is this information real or fabricated?'

Taiwan's FactCheck Center recently exposed the serious impact these videos have on the information environment. Editor-in-Chief Chen Weiting told BBC Chinese that YouTube channels operated by content farms have surged, often featuring virtual presenters claiming to be 'doctors,' 'retired Taiwan Railway employees,' or 'former flight attendants.' These videos not only deliver health advice targeting seniors under the guise of medical authority but also promote false 'senior welfare policies.' For example, they claim high-speed rail tickets automatically upgrade to business class, companions ride free, or the government will distribute cash or provide free nursing care during hospitalization—leading many seniors to inquire at stations, causing confusion. Both Taiwan Railway and Taiwan High Speed Rail have issued multiple statements this year clarifying that so-called 'senior discount programs' or 'hidden benefits' are entirely fictional.

Specifically Targeting Middle-Aged and Elderly Audiences

Most seniors in the class were regular viewers of such AI videos. One participant admitted visiting the high-speed rail station hoping to receive the advertised benefits, only to be disappointed. Another, 86-year-old Wang Fumei, said she used to fully trust information forwarded by friends and frequently shared it. Now, realizing these were AI-generated 'fake videos,' she feels 'deceived.' 'I’ve started doubting people. In the AI era, life won’t be very pleasant,' she told BBC Chinese.

While these videos on YouTube are often labeled as AI-generated, most seniors don’t notice. Google, YouTube’s parent company, stated that its platform rules apply equally to AI-generated content and prohibit medically false information that could cause serious real-world harm. Dai Yuchen, a researcher at Taiwan’s Information Security Research Institute, told BBC Chinese that these videos deeply impact daily life in Taiwan. Among over 200 YouTube channels he tracked, more than 70,000 videos were uploaded, accumulating 1.32 billion views—equivalent to every person in Taiwan watching about 60 times on average. He analyzed that nearly all these videos are AI-generated, produced at extremely low cost and in massive volume. Typically, AI constructs the 'doctor’s' face and background, then synchronizes lip movements with audio. 'With the latest tools, a 20-second video can be completed in seconds,' he said. The content is often automatically scraped and stitched together from online sources by AI, with subtitles added automatically. Creators need no expertise or editing skills to produce seemingly professional videos quickly.

Dai emphasized that these channels focus on medical wellness, retirement life, science education, and religion—'specifically targeting middle-aged and elderly audiences.' Because the characters appear realistic and the content resonates with daily concerns, viewers easily believe the on-screen figures are real. Community college teacher Li Ziqiang added that these videos gain trust by using authoritative 'experts' and cater directly to seniors’ interests. 'Health is their top concern, and financial incentives like discounts or transportation deals make them especially attentive.'

Fulfilling Elderly Psychological Needs

He Yiqing never imagined he’d become a victim of misinformation. A former journalist, he enjoys learning and sharing information even in retirement, spending over five hours daily watching and forwarding videos shared via messaging apps. His content covers health, life knowledge, and social issues. Before bed, he categorizes them, then spends one to two hours each morning forwarding them. This routine has lasted three years, becoming a fixed part of his life. 'It keeps my mind active, helps me learn, and connect with friends,' He said. He forwards to at least 50 friends at once. Though fewer than half respond, it gives him a sense of connection. In recent years, caring for grandchildren reduced his social and ballroom dancing activities, leaving more free time.

'Some AI videos are obviously fake, like parody clips mimicking Putin or Trump,' he said. Since the course, he checks video authenticity and warns group members: 'This isn’t real.'

Dr. Yuan Wei, attending physician in psychiatry at Taipei Zhenxing Hospital and specialist in geriatric mental health, told BBC Chinese that many AI health videos are designed to exploit seniors’ 'anxiety.' As age brings accumulating health issues, people feel a loss of control. 'These videos offer actionable solutions, making seniors feel they can do something, suppressing anxiety and restoring a sense of control.'

Dr. Yuan added that sharing videos is also a form of social connection. Seniors may face reduced planning ability and environmental interaction due to physical conditions or mild cognitive decline. In contrast, using smartphones is low-barrier, allowing them to express care. Positive responses bring joy, a sense of being 'seen,' and even the feeling of helping others.

Li Ziqiang added that for many seniors, watching AI videos provides companionship and fills time. Unlike younger users who prefer short videos, seniors often watch videos over 20 minutes long, leaving them on autoplay like a 'radio' to fill daily voids.

Political Motives?

As these videos’ influence grows, scrutiny over their origins intensifies. Researcher Dai Yuchen noted that most monitored YouTube channels obscure their operators’ identities, with registration locations spanning Taiwan and various countries. Some involve clusters of seven or eight bot accounts with nearly identical profiles. He said these channels primarily generate traffic for ad revenue, but their ultimate purpose remains unclear. Past research shows many traditional online content farms are run by 'cyber armies' that can inject political content when needed. 'Many of these channels are high-risk for me. They need monitoring—once they shift style, algorithmic promotion can reach vast elderly audiences,' Dai added.

In 2024, Taiwanese political scholar Wang Hong’en published research showing a Chinese company controlled multiple Hong Kong accounts, creating non-political lifestyle and wellness Facebook pages to attract traffic, then inserting political propaganda and cognitive warfare content during Taiwan’s presidential election. BBC Chinese also investigated 20 high-subscription channels but found none disclosed operator names. Creation times varied, with locations listed in Taiwan, the U.S., Malaysia, and Hong Kong. Some channels were removed for violating platform policies.

An anonymous Taiwanese national security source told BBC Chinese that over the past year, Taiwan observed several such AI video channels shift to supporting specific candidates during elections. With local elections approaching, a few channels have recently begun spreading messages like 'unification with China is better.' In response to concerns about Chinese interference and cognitive warfare, China’s Taiwan Affairs Office repeatedly denied involvement, claiming the DPP government is the true 'cognitive warfare' manipulator. Chen Huimin, founder of FactLink Digital Literacy Lab, told BBC Chinese that these videos are now highly scrutinized, with suspicions of coordinated manipulation—especially critical as elections near. She noted that even without evidence of malicious intent, these videos qualify as 'AI slop'—poorly made, low-information-content content produced at scale.