Editor's Note

Last month, a doctor friend shared his confusion with me.

A patient had just sat down when he thrust his phone in front of the doctor. The screen was densely packed with AI-generated "treatment recommendations"—from medication plans to lifestyle restrictions, covering every detail. With a hint of scrutiny, the patient asked: "Doctor, the AI says this medication has serious side effects. Your prescription is different. Could you double-check?"

My friend was stunned. He had encountered patients who came with Baidu search results to "argue their case," but this was the first time a patient had brought an AI-generated "prescription" to "guide" his prescribing. And this is not an isolated incident.

By 2026, AI is no longer a concept confined to laboratories. It has entered ICUs, tracing entire diagnostic and treatment journeys in 5 seconds. It has entered operating rooms, assisting screw placement with 0.1mm precision. It has even entered patients' smartphones, becoming a "digital doctor" accessible to everyone.

So, an unavoidable question confronts all medical professionals: Is AI a powerful assistant that adds wings to the tiger, or a rival trying to snatch food from the tiger's mouth?

01. When AI Enters the Consultation Room

First, let's talk about a recent internet meme—everyone is "raising lobsters." The "lobster" here isn't a delicacy on the dinner table, but an AI agent called OpenClaw. Its icon resembles a lobster, so netizens call the process of deploying and training it "raising lobsters." This AI, which can act as a "digital twin," autonomously perform computer operations, and handle various tasks, has shown people the high efficiency value of intelligent tools, and this craze has also extended to the medical field.

Image Source: Doubao

While people are enthusiastic about training AI to handle daily affairs, AI in the medical field has already left the lab and begun tackling tasks hundreds of times more complex than "booking a flight"—saving lives.

In the ICU of Peking University Shenzhen Hospital, AI can trace back all data from the entire diagnostic and treatment process within 5 seconds and generate a structured medical record in 1 minute. In surgery, AI 3D reconstruction technology converts lung CT scans into 3D models, clearly showing tumor location, invasion range, and relationship with surrounding tissues, transforming "evaluative gross anatomy" into "navigated fine anatomy," helping surgeons maximize the preservation of healthy lung tissue.【1】 In the past, such work would keep a doctor busy for half a day.

Data from Xi'an North Hospital shows that AI has reduced the imaging diagnosis time for aortic dissection from 15-20 minutes to just 3 minutes, buying precious time for life-saving interventions. In detecting early breast cancer and occult rib fractures, AI can mark suspicious lesions in seconds, helping doctors reduce the risk of missed diagnoses.【3】

According to information released by the National Health Commission on November 27, 2025, 80% of counties (cities/districts) in China have initially established county-level shared resource centers for imaging, electrocardiograms, and laboratory tests. In 2025, county-level remote medical imaging diagnostic services exceeded 68 million person-times, and over 360 cities have achieved mutual recognition of more than 200 types of examination and test results, steadily advancing the "examination at the grassroots, diagnosis at the district/county level, mutual recognition of results" model.【1】

AI is no longer a future concept; it is the present continuous tense of the medical field in 2026. However, as the technological wave crashes ashore, the first to feel it are often not the direction of the wind, but the footprints on the beach wet by the waves.

02. Doctors' "AI Anxiety"

In February this year, a post on a medical forum exploded with discussion.

A young doctor wrote: "I studied hard for 8 years, just entered the department, and found that AI reads images faster and more accurately. I can't help but ask myself: What exactly did I learn in these 8 years?" The comment section erupted. Some comforted him: "AI is just a tool, don't panic." Others delivered a blow: "Radiology departments might not need so many people in the future."

This is not alarmist. Repetitive, standardized, memory-based tasks—AI is indeed better than humans. A discharge summary takes a doctor 10 minutes; AI does it in tens of seconds. Quality control of images—a doctor gets tired after looking at hundreds a day; AI doesn't. Patient inquiries—AI can respond 24/7. Some roles will indeed be restructured.

However, deeper than "job security anxiety" is the动摇 of the trust relationship. When AI's suggestions contradict a doctor's judgment, who should the patient trust? As in the opening scenario, AI only sees the side effects listed in the drug insert, but cannot see the patient's specific situation—blood pressure, heart rate, medical history, drug tolerance… But patients don't understand this. They only see that what the AI said is different from what the doctor said.

This is the real source of anxiety: AI is changing the trust relationship between doctors and patients.

03. Three Hurdles AI Cannot Cross

So, will AI replace doctors? The answer is no. Because it can never overcome three core hurdles.

First, AI lacks "clinical empathy."

Medicine is never just "reading images + prescribing medication." When a patient is anxious, a doctor's reassuring "Don't worry" or a confident look can be as effective as half a dose of medicine. AI can never understand what "fear" feels like. A patient brought an AI-generated Traditional Chinese Medicine prescription to a doctor. The prescription itself was correct, but considering the patient's constitution, lifestyle habits, and recent condition, adjustments were needed—some herbs reduced, others substituted. AI couldn't do that. It can only provide a "standard answer" based on data.

Second, AI cannot handle "complex conditions."

AI models are trained on existing data. For rare diseases or complex cases involving multiple intertwined conditions, AI's recommendations often have low accuracy. Doctors, drawing on years of clinical experience, can make flexible judgments based on the patient's specific symptoms and history, even breaking new ground beyond existing guidelines to achieve personalized diagnosis and treatment. Hu Min, head of the Information Center at Xi'an North Hospital, stated: "At this stage, AI's main value is 'speeding up + reducing missed diagnoses,' not surpassing human judgment in all scenarios."【5】

Third, AI cannot take responsibility.

This is the most crucial point. Chinese laws and regulations clearly stipulate that AI does not have prescribing authority. In case of a medical dispute, the responsibility ultimately falls on the doctor. AI can give suggestions, but the final decision must be made by a doctor. Only a doctor can take responsibility for that decision.

04. A New Consensus is Forming in the 2026 Medical Community

Driven by both top-down policies and bottom-up clinical needs, as large-scale AI implementation in healthcare progresses in 2026, the industry has moved beyond debating "whether AI will replace doctors" and reached a general consensus: AI is not a competitor to doctors, but an "amplifier of doctors' capabilities."

The "15th Five-Year Plan for National Economic and Social Development (Draft)"摘要 proposes the comprehensive implementation of an "artificial intelligence +" action. As AI deeply integrates into medical care, what will future hospitals look like? How will the doctor's role transform? How can high-quality medical resources truly reach the grassroots? These questions are moving from speculation to reality.

During the National Two Sessions, Yang Zhenglin, a National People's Congress representative, academician of the Chinese Academy of Sciences, and Party Secretary of Sichuan Provincial People's Hospital, told a Beijing News reporter: "Doctors who don't embrace AI may find themselves in a very difficult position in the future," offering his profound insight into the upcoming medical transformation.【7】

Under this consensus, AI plays different roles for doctors at different stages:

For senior doctors, AI is an "efficiency multiplier."

Senior doctors often handle a large volume of complex cases and research. AI helps them quickly handle tedious paperwork, initial image screening, etc., freeing up more time for case analysis, research innovation, and clinical teaching. For example, with AI, doctors can rapidly search vast medical literature, track guideline changes, and even analyze departmental clinical data to identify research entry points, advancing medical technology.

For junior doctors, AI is a "growth accelerator."

Junior doctors lack clinical experience. AI can provide them with standardized diagnostic recommendations and references for similar cases, helping them quickly familiarize themselves with diagnostic standards and improve their abilities.

For grassroots doctors, AI is a "capability equalizer."

County hospitals and community health centers lack experts the most. Now, AI-assisted decision-making systems can help general practitioners standardize their consultation process and provide diagnostic, medication, and referral suggestions. Grassroots patients don't need to travel to big cities to receive care close to that of top-tier hospitals.

This is AI's true value: making scarce, high-quality medical resources replicable and accessible at the grassroots level.

05. The Essence of "Healers" in the AI Era

How should medical professionals position themselves in the AI wave?

First, cultivate AI literacy.

This doesn't mean doctors should learn to code, but rather understand: What can AI do? What can't it do? When to leverage it, and when to question it? When AI suggestions differ from a doctor's experience, ask "why?" and take an extra step to verify. This critical thinking is precisely the irreplaceable value of doctors.

Second, participate in AI iteration.

The best medical AI is born from clinical practice. If you can tell engineers, "this feature isn't practical" or "there's a flaw in that logic," you are participating in defining the future of medical tools.

Third, embrace lifelong learning.

The more powerful the tools, the more valuable human judgment becomes. AI can help doctors quickly master new knowledge, but the final judgment, integration, and application still rely on the doctor's wisdom.

In 2026, AI is indeed changing medicine. It makes diagnosis faster, surgery more precise, and frees doctors from tedious administrative tasks. But it cannot change the essence of medicine.

The essence of medicine is the human-to-human connection. It's the doctor's hand holding the anxious patient's. It's the doctor's words, "Let's try again," to the desperate family member. It's the difficult decision made together by doctor and patient in the face of life and death. These are things AI can never do.

So, back to the opening question: Is AI an assistant or a rival? The answer is clear—AI is an assistant. And doctors will always be doctors.

Efung Knowledge Planet will continue to follow the latest developments in biomedicine. In the AI healthcare, Yifeng has initiated systematic, accompanying technologies from the lab to the clinic, witnessing the best collaboration between "assistant" and "healer." If you are a medical professional, welcome to share in the comments: What changes has your AI tool brought to you?

References:

1.https://mp.weixin.qq.com/s/dSY8-aOoVj241sqfGpNHRQ

2.https://mp.weixin.qq.com/s/GWJ_78KRBWy2YN5-o1h7hA

3.https://mp.weixin.qq.com/s/s45JLpsVY95-GNYvQZtQMg

4.https://mp.weixin.qq.com/s/rOIbkLb8lC8UdiwGdepGEA

5.https://mp.weixin.qq.com/s/rEcJvFRcb8ksToKeW-lr9w

6.https://mp.weixin.qq.com/s/x7dBkYMr4tb5LYj8y2uQzA

7.https://mp.weixin.qq.com/s/Gd6Jtafv4NDetouONAn4VQ