Starting in October 2025, anyone in China who wants to post about medicine, finance, law, or education on Douyin, Weibo, or Bilibili has to show a degree first. The Cyberspace Administration of China made the rule. Creators who cannot verify a relevant qualification can be suspended, and individual fines reach roughly 100,000 yuan, about $13,000.
The stated reason is misinformation. Researchers studying social platforms have repeatedly found that health misinformation spreads faster and farther than careful health information, and that this spread changes what people actually do.1,2 The Chinese rule is one government’s answer to that pattern. Whether it is a good answer is the more interesting question.
What the new Chinese rule actually says
The policy applies to four categories the regulator calls sensitive: medicine, finance, law, and education. A creator posting in any of those areas has to submit verified degrees or professional certifications to the platform before publishing. Without that paperwork, the account can be limited or suspended, and individual fines can reach 100,000 yuan. The platforms named in coverage so far include Douyin (the Chinese version of TikTok), Weibo, and Bilibili, which together reach hundreds of millions of daily users.
The framing in state media is straightforward. A doctor talking about hypertension drugs should be a doctor. A lawyer explaining a divorce statute should hold a license. A finance creator pushing a stock pick should be a registered analyst. The rule does not, on its face, ban discussion of any topic. It binds public claims in those four areas to a paper credential.
Outside China, reactions split predictably. Press-freedom groups read the rule as another tool to silence inconvenient voices in a system that already polices speech heavily. Public-health researchers, who watch dangerous medical claims rack up millions of views every week, were quieter and, in some cases, curious. The reason is simple. The research on health misinformation is not flattering.
How big is the misinformation problem, really?
A 2021 systematic review in the Journal of Medical Internet Research, led by Suarez-Lledo and Alvarez-Galvez, pooled 69 studies of health misinformation on social media.1 The patterns were grim and consistent. False or misleading content was common across topics like vaccines, smoking, drugs of abuse, eating disorders, pandemics, and noncommunicable diseases. Misinformation was especially heavy in posts about smoking products and drugs, where commercial interest meets thin moderation. Some platforms acted as larger amplifiers than others, and certain topics, including vaccines, drew unusually high rates of false content.
The deeper finding was about reach. In a 2018 paper in Science, Vosoughi, Roy, and Aral analyzed roughly 126,000 cascades of news stories on Twitter, spread by about 3 million people from 2006 to 2017.2 False stories diffused significantly faster, deeper, and more broadly than true ones. The top 1% of false-news cascades reached between 1,000 and 100,000 people; true stories rarely passed 1,000. Falsehood was about 70% more likely to be retweeted than truth, and the effect was strongest for political news, but it held for medical and scientific topics too. The drivers were human, not bot. People shared false stories more, the authors argued, because falsehood tends to be more novel.
That is the structural problem any country is reacting to when it touches platforms. False content is not just present. It travels. And it travels through ordinary people, not shadowy actors. The Vosoughi team specifically tested whether bots were responsible for the asymmetry, and the answer was no. When automated accounts were filtered out of the data, the gap between the reach of false stories and true stories barely moved. The amplifier was us.
Does misinformation actually change behavior?
It is one thing for a wrong claim to circulate. It is another for it to nudge a real decision in the offline world. The honest answer from the literature is that misinformation does change behavior, modestly but measurably, in the kinds of decisions that matter for public health.
A 2021 study in Nature Human Behaviour by Loomba and colleagues at the Vaccine Confidence Project ran a randomized experiment in the UK and the USA just before the COVID-19 vaccines rolled out.4 Participants were shown either factual information about the vaccines or a curated stream of misinformation that was already circulating online. Exposure to misinformation reduced people’s stated intent to accept a vaccine for themselves by about 6.2 percentage points in the UK and 6.4 percentage points in the USA, on top of an already shaky baseline. Reductions were larger when participants were asked about vaccinating others. Some demographic groups, including people with the lowest baseline confidence, were affected more.
Six points sounds small until you scale it across a country. Vaccine programs that need 80% or more uptake to blunt transmission do not have a six-point cushion to spare. And the misinformation in that experiment was not exotic; it was the kind of post that already circulated on Facebook, Instagram, and forwarded WhatsApp messages during the 2020 to 2021 winter. People in the study saw the same kind of content their relatives were sending them.
The behavior research has another wrinkle worth knowing. In a series of experiments published in Psychological Science in 2020, Pennycook and colleagues showed that people often share inaccurate health content not because they believe it, but because their attention is on social signals rather than truth.3 A simple nudge that reminded participants to think about whether a headline was accurate before sharing improved the quality of what they passed along. People are not necessarily fooled. They are distracted, and distracted sharing fills timelines with noise.
Where credentials help, and where they do not
The case for asking credentialed people to identify themselves is intuitive. A urologist talking about prostate cancer is more likely than a wellness coach to know which interventions extend life and which do not. In a 2019 study in European Urology, Loeb and colleagues evaluated the 150 most-viewed YouTube videos about prostate cancer.5 A panel of experts found misinformation in 77% of the videos, and biased or commercially motivated content in many of the rest. The total view count of the videos studied was over 6 million. The patients watching them were often making real decisions about screening and treatment.
That is the kind of failure a credential rule could plausibly catch. If only board-certified urologists, oncologists, or licensed clinicians could post a claim about a prostate-cancer protocol, the worst videos in that 77% might never go up.
The case against is less about credentials and more about who decides which credentials count, and what the rule covers when it gets pushed past its original target. A formal degree does not always equal accurate or helpful content. Some of the most useful health communicators online are nurses, dietitians, physician assistants, peer counselors, or patients with hard-won lived experience. Some of the worst are doctors with mail-order supplement lines. The history of medical paternalism, from lobotomy enthusiasm to early HIV stigma, is also a history of credentialed people being confidently wrong while the patient community got there first.
The other risk is a rule that drifts. A regulation written for medicine and finance can be widened, in a system without independent courts, to cover nutrition, parenting, sex education, or anything else a sitting government finds inconvenient. The line between a quack post and a dissenting one is sometimes obvious, sometimes not, and the body that draws it has enormous power.
What other countries are doing
China is far from alone in trying to put a hand on the platform-health problem. The European Union’s Digital Services Act, in force from 2024, requires very large online platforms to assess and mitigate systemic risks, including risks to public health from misinformation. Enforcement is administrative, not criminal, and the rule does not single out individual creators by credential.
The United Kingdom’s Online Safety Act, also taking effect through 2024 and 2025, focuses more on illegal content and harms to children, and uses platform duties rather than creator licensing. In the United States, federal courts have repeatedly cited the First Amendment to block direct restrictions on health speech, leaving most enforcement to platforms themselves and to professional licensing boards that can discipline doctors who post falsehoods under their medical license.
Australia, Canada, and several Latin American countries are running consultations about health misinformation, but none have yet adopted a credential gate that looks like China’s. Most are inching toward platform-level transparency, takedown speed, and labeled corrections, not personal certification. India has used emergency takedown orders during specific public-health events, including dengue and COVID-19 surges, but stops short of a standing credential rule for creators.
So there is a spectrum. On one end, China’s hard credential rule with personal fines. On the other, lighter regimes that ask platforms to act without dictating who is allowed to speak. Researchers can tell you what the misinformation problem looks like. The choice between these regimes is mostly political.
Common questions about the new rule
Does the Chinese rule ban anyone from talking about health?
No. It binds public posts in four sensitive areas, including medicine, to a verified credential check on the platform. People without those credentials can still post about food, lifestyle, fitness, beauty, or general wellbeing. The line between general wellness and medical advice will, in practice, be drawn by platform reviewers.
How big are the fines, exactly?
Coverage from the Cyberspace Administration of China and outlets including BBC News, Reuters, and South China Morning Post puts the maximum individual fine at about 100,000 yuan, which is roughly $13,000 at recent exchange rates. The source post quotes the same number. Platforms can also face their own penalties for failing to verify creators.
Is there evidence that credential gates reduce health misinformation?
The direct evidence is thin because the policy is new. Indirectly, studies of doctor-led versus layperson-led health content tend to find that licensed clinicians produce content with fewer factual errors on average, although biases and commercial conflicts still appear.5 More carefully designed evaluations of the Chinese rule will not be available for a year or two.
Could a similar rule work in democracies?
Constitutional speech protections in the United States make a Chinese-style rule unlikely there. The European model puts duties on platforms rather than on individual creators, which avoids some censorship risks at the cost of being less precise. Democracies will likely keep mixing platform duties, professional licensing, and labeled fact-checks rather than introducing personal credential gates.
What can a reader do today?
Three habits help. Check whether the person making a health claim has any relevant training listed in a verifiable place. Look for whether the claim cites a study with a real journal and a date. Pause before sharing. The accuracy nudge from the Pennycook study works in the lab, and it works on a phone too.3
Reading the bigger pattern
The Chinese rule is not a clean experiment. It is one country’s strict, top-down attempt to fix a problem that every country is struggling with. The research is unambiguous that health misinformation is widespread on social platforms, that it spreads faster than accurate content, and that it changes real-world behavior, including vaccination intent.1,2,4 Doing nothing has costs. Doing the wrong thing also has costs.
What is worth keeping in view, as more governments start to legislate, is that the goal is not silence. The goal is fewer people taking medically wrong actions because of a confident video. There are several routes to that goal. Better platform design, including accuracy nudges and source labels. Stronger licensing-board enforcement, where the credentials people already hold mean something when they speak in public. Reader habits that pause before sharing. And, in some systems, the kind of credential gate China just built. The trade-offs differ by country and by political tradition. The underlying problem does not.
Sources
- Suarez-Lledo V, Alvarez-Galvez J. Prevalence of Health Misinformation on Social Media: Systematic Review. Journal of Medical Internet Research. 2021. PubMed: 33470931
- Vosoughi S, Roy D, Aral S. The spread of true and false news online. Science. 2018. PubMed: 29590045
- Pennycook G, McPhetres J, Zhang Y, Lu JG, Rand DG. Fighting COVID-19 Misinformation on Social Media: Experimental Evidence for a Scalable Accuracy-Nudge Intervention. Psychological Science. 2020. PubMed: 32603243
- Loomba S, de Figueiredo A, Piatek SJ, de Graaf K, Larson HJ. Measuring the impact of COVID-19 vaccine misinformation on vaccination intent in the UK and USA. Nature Human Behaviour. 2021. PubMed: 33547453
- Loeb S, Sengupta S, Butaney M, et al. Dissemination of Misinformative and Biased Information about Prostate Cancer on YouTube. European Urology. 2019. PubMed: 30502104
