AI health consultations such as Doubao's gain popularity: why are big tech companies redoing internet healthcare?

In 2026, when doctors sit buried in consultations, they expect to be facing ordinary patients seeking care—but what they get instead is a crowd of "AI medical expert" fact-checkers.

Recently, a notice created by netizens as a meme went viral and hit the top of trending searches.

According to a report by Yicai, Doubao's daily health consultation volume has surpassed 20 million. Public data from the National Health Commission shows that the total daily offline outpatient volume across all public and private hospitals nationwide is just over 10 million.

A single Doubao is doing double the work of 5 million doctors.

It's not just China's AI sector that's booming. Official OpenAI data disclosed in January 2026 shows that over 230 million people worldwide ask health-related questions on ChatGPT each week, and by July of this year that figure had soared to 300 million.

And it's not just young people in big cities using AI for medical care—more than half of Ant Afu's 150 million users come from third-tier cities and below.

Rewind ten years, and the standard move for Chinese people before seeing a doctor was to "Baidu it first."

That generation of online medical care was essentially piecing together a diagnosis from a jumble of disorganized web information. Quality varied wildly and the logic jumped around, but because it looked "unofficial," most people still ended up going to the hospital.

Headache = brain tumor, cough = lung cancer, stomachache = pancreatic cancer. Someone bites the inside of their mouth while eating, does a casual search, and the result is cancer.

Any minor symptom can produce the feeling of a terminal illness—the more you search, the more panicked you get, and the more panicked, the more anxious, making it truly "self-diagnosis equals terminal illness."

More importantly, it has a sense of proportion—it will remind you that if your hands are numb or weak, or you have dizziness and headaches, don't sit there slowly asking your phone questions; get to the hospital right away.

The cost of in-person medical care isn't just registration and medication fees.

If you feel unwell and want a consultation, you need to take leave, wait in line, sit in traffic, and spend half a day—and the time actually talking to the doctor may be less than three minutes. For working people, going through all that hassle for a minor ailment has an extremely low cost-benefit ratio.

A 2025 survey of 23 top-tier hospitals nationwide showed that the average consultation time per patient in general outpatient clinics in China is only 4.8 minutes, and even for associate chief physician and above specialist clinics it's only 7.6 minutes.

In other words, in those few minutes the doctor has to complete the entire process, which leaves very little time for patients to keep asking questions—tight time, heavy workload, high pressure on doctors, and patients are nervous too.

So many people feel awkward about asking the doctor questions, and don't want to make a special trip to the hospital, so they just quietly tough it out.

A child has a fever at dawn—should you use physical cooling or give medicine right away? An elderly person has abdominal pain in the middle of the night—is it ordinary stomach upset or does it need emergency care? Are the arrows on the physical exam report actually serious? Which department should I register with for a follow-up? Can these two medications be taken together?

In October of this year, a netizen in Guizhou was hospitalized for abdominal pain, but after completing a full set of tests in the gastroenterology department, no cause could be found. Sensing something was wrong, he contacted his ex-wife, hoping she would help him draft a will in case the worst happened.

Another netizen shared: "Even though many people criticize Doubao, it really saved my dad's life. That night his tongue suddenly went numb and he couldn't lift his legs. I figured it was nothing serious and we'd go to the hospital in the morning. But I searched on Doubao and it said this was a cerebral infarction! We went to the hospital that very night and got rapid thrombolytic treatment, preserving the golden treatment window. I'm still grateful to Doubao."

QuestMobile 2026 industry data shows that monthly active users of native AI applications in China have reached 440 million, with health consultation being the most frequently used scenario, far exceeding office work, study, and other functions.

Why are the big tech companies rushing into what was once a loss-making "hard business"?

Back in 2015, the first internet hospital appeared, and subsequently internet healthcare and online consultations sprang up like bamboo shoots after rain.

Because medical resources are very unevenly distributed, seeing a doctor at a well-known top-tier hospital meant queuing for a number, buying numbers from scalpers. In 2014, Chunyu Doctor raised $50 million in Series C funding, setting the record at the time for the largest single financing in the mobile health field, and its founder proposed to build "the Alibaba of the mobile health field."

But later, these internet healthcare platforms did not fare well:

Chunyu Doctor lost money all the way and ultimately ended up "being sold"; Weiyi accumulated losses of over 8 billion yuan from 2021 to the first half of 2024, with a bumpy road to profitability; Haodf Online was acquired by Ant Group.

But now, from ByteDance's Doubao and Ant Afu to DeepSeek and ChatGPT, all the leading AI players treat health Q&A as a deployment scenario.

Why are the big tech companies fighting to enter what was once a loss-making hard business?

The collapse of Chunyu Doctor, Weiyi and their peers back then came down to an unavoidable "heavy-asset dead end."

The old internet healthcare players essentially just moved offline clinics online, with no change to the core. For a platform to operate, it had to sign up massive numbers of licensed doctors, build customer service and operations teams, and handle consultation orders—whether labor costs or operating costs, all were rigid expenditures, and the more users, the higher the costs.

To make matters worse, under China's medical regulatory rules, traditional internet hospitals can only handle follow-up visits, not first visits—they only serve people with chronic disease checkups and prescription renewals.

This left the industry in an awkward predicament—the services the public truly needed couldn't be provided, and the services that could be provided weren't used frequently by anyone, ultimately reducing them to "niche, low-frequency tools."

What looks like a comeback in the same sector is actually technological iteration completely rewriting the fate of medical services.

In January of this year, Life Times published a survey involving over 500 top-tier hospital doctors nationwide.

The survey results show that over 70% of top-tier doctors said they would be willing to recommend that the public use AI doctors to address everyday basic health questions, 62% of the doctors surveyed are already using AI doctors to assist their own work, and 90% of the doctors surveyed said they are very optimistic about the future development of AI doctors.

The core reason is:

General-purpose AI large models represented by Doubao, DeepSeek, and ChatGPT are essentially language models, not professional medical systems.

A language model's core capability is integrating, summarizing, organizing, and outputting standardized textual information. It has read vast amounts of medical textbooks and popular science content, but it has no real clinical experience, no in-person diagnostic perception, and certainly not the clinical intuition and case judgment that doctors accumulate over years.

A study published by scholars from Hangzhou Normal University and Hangzhou Medical College pointed out that 11% of people over 45 who have an illness do not seek medical care, with the most common reason being "there's no need to see a doctor."

So the first hurdle in medical treatment often appears before the hospital.

Media reports say that in Dalian, a 75-year-old man named Zhou Mingyuan was awakened by pain late at night. The first time AI alerted him that he might be having an acute myocardial infarction and should call 120 immediately, he didn't call emergency services because he thought his health had always been good—it wasn't until the AI urged him a second time that he must call 120 that he took it seriously, and he ultimately had surgery that night and successfully saved himself.

For example, Ant Afu has announced that Afu sends out more than 200,000 "seek medical care immediately" recommendations every day, over 50,000 of which are related to critical conditions.

So the value of AI is not to replace doctors, but to serve as a front-line gap-filling tool, a tireless, 24-hour online assistant.

This article is from the WeChat public account "Jincuodao" (ID: ijincuodao), author: Zhang Yichi, published by 36Kr with authorization.