DeepSeek涉足医疗：AI与医生的博弈

DeepSeek在医疗领域引发的一系列现象，包括患者使用DeepSeek看病后对医生治疗方案产生质疑、医生分享类似经历，还介绍了多家医院接入DeepSeek的情况，分析了AI在医疗领域的优势、短板，阐述了医疗领域中可以应用AI的场景，探讨了用DeepSeek看病带来的诸多问题以及其在医疗领域进一步发挥作用的可能和需要关注的问题等。

The "big model heat wave" triggered by DeepSeek continues. Some people use it for programming, some for fortune - telling, and others for medical consultation.

There was a very interesting news these two days. A medical blogger posted, "The sky is falling! After the patient used DeepSeek, they questioned my treatment plan. I was so angry that I checked the guidelines again myself, and only then did I find that the guidelines had been updated..." As of 12:00 today (the 26th), this post has received 154,000 likes and nearly 30,000 comments.

Similar news is not an isolated case. Doctors in Beijing, Shanghai, and Hangzhou have all shared similar experiences. In the comment section, some netizens also said that they had encountered the same situation. One netizen said: "Yesterday during the outpatient service, after I finished explaining to the patient. He showed me the results checked by DeepSeek, which were the same as what I said, and the patient praised me for being professional and responsible." This netizen joked, "I feel that I'm not far from being laid off."

AI has stirred up a thousand waves in the medical field and triggered heated social discussions. Many physicians have claimed that "DeepSeek's diagnosis results are astonishing and have sounded the alarm for doctors." Is it reliable to use DeepSeek for medical treatment and prescribe prescriptions? Whose lives are being saved when using DeepSeek for medical treatment? How to correctly use DeepSeek in the medical field has become a difficult problem faced by both doctors and patients. Should we trust AI or doctors? Will AI diagnosis and treatment replace doctors?

01

DeepSeek Diagnosis: Sounding the Alarm for Doctors

Under the AI boom triggered by DeepSeek, the medical field is inevitably affected.

According to incomplete media statistics, since February, more than 30 hospitals across the country, such as Tianjin Medical University General Hospital, Shenzhen University Affiliated South China Hospital, and Shanghai Sixth People's Hospital, have announced their access to DeepSeek.

First of all, it must be affirmed that the development prospects of AI in medical care cannot be underestimated. Especially in medical scenarios with a high degree of standardization such as image recognition and auxiliary diagnosis, the advantages of artificial intelligence are more obvious.

The fact that patients can think of using DeepSeek for medical treatment shows that patients have the will and need to "acquire medical knowledge", which is not entirely a bad thing.

On February 21, Dr. Shen Ling, chief physician of the Respiratory Department of Hangzhou First People's Hospital, shared his recent medical experience: A female patient was diagnosed with post - nasal drip syndrome, but during the more than two - year medical treatment process of this patient before, none of the doctors had mentioned post - nasal drip syndrome. Dr. Shen's diagnosis was consistent with the judgment given by DeepSeek, which surprisingly made the patient full of joy. Dr. Shen Ling posted on his personal social media account to remind his colleagues: Patients' "self - diagnosis" with AI has sounded the alarm for doctors!

02

What Has Changed After Connecting to DeepSeek?

For patients, when AI tools have been integrated into daily life and work, it is natural to use AI for health consultation, film reading, and medical consultation.

It must be admitted that AI connected to the medical knowledge base has broken the medical information barrier and can improve the accuracy of diagnosis. Especially in areas with a high degree of standardization such as rare disease identification and image recognition, AI can show significant advantages.

However, it should also be noted that when AI is widely used in the medical field, doctors will find it difficult to adapt in the short term. In the past, when patients searched on the Internet before seeing a doctor, some doctors were already in a difficult situation: Whether they respected themselves was secondary. The main problem was that the information on the Internet was mixed, which not only increased patients' anxiety but also made doctors pay more explanation costs. And now, in front of DeepSeek, doctors are facing greater pressure. When AI provides the "standard answer" for patients, while what the doctor provides may be outdated or even wrong information, the resulting embarrassment can be imagined.

In many aspects, AI also has irreparable shortcomings. For example, as a key part of the diagnosis and treatment process, prescription writing has irreplaceable characteristics in terms of professionalism and responsibility. Physicians need to base their prescription writing not only on professional medical knowledge but also on comprehensive judgment by combining multi - dimensional information such as patients' individual differences, medication history, and allergy history. Even if AI is trained with massive data, it is difficult to fully simulate the complex decision - making process in clinical diagnosis and treatment, and it is even more unable to bear the main responsibility for medical actions.

03

Which Areas of Medical Care Can AI Be Applied to?

At the end of 2024, the National Health Commission issued the "Reference Guidelines for Artificial Intelligence Application Scenarios in the Health Industry" to actively promote the application and innovative development of "artificial intelligence" in the health industry, which clearly defined 84 typical application scenarios in 13 application areas of artificial intelligence in the medical and health industry.

In terms of "medical services", it is proposed that AI can be applied in areas such as intelligent auxiliary diagnosis of medical images, intelligent auxiliary quality control of medical image data, intelligent auxiliary decision - making for clinical specialized diseases, intelligent auxiliary decision - making for general practitioners at the grass - roots level, intelligent auxiliary treatment of medical images, intelligent auxiliary planning for surgery, and intelligent auxiliary delineation of radiotherapy target areas.

It can be seen that in non - core medical links such as medical consultation, health management, and medical record analysis, artificial intelligence can indeed effectively improve efficiency. At present, the application of AI technology in the medical field still plays a supplementary or auxiliary role in terms of means or the choice of diagnosis and treatment plans, and the final decision is still made by doctors or relevant medical technicians.

04

Using DeepSeek for Medical Treatment: Whose Lives Are Being Saved?

Medical AI is not a new thing. However, recently DeepSeek has become very popular and has broken out of the circle in a turbulent manner, conducting "AI use education" for doctors and patients, making more doctors embrace AI technology and making patients trust it enough to use AI for health consultation, film reading, and medical consultation.

Different from consultation dialogues in other life scenarios, medical consultation and cancer screening require high accuracy and have low tolerance for errors. To cope with the new challenges brought by artificial intelligence to the medical field, it is more important to correct people's positions. Medical workers should upgrade their technological awareness, learn about the reconstruction of work processes by AI, and adapt to the "human - machine double - review" model, and gradually shift from being a simple technology executor to a clinical strategy maker. Patients should also have a clear understanding. Not every patient can get a scientific and comprehensive answer. The completeness and professionalism of patients' questions also determine the accuracy of DeepSeek's answers. Do not let data deviation and algorithm traps turn into medical damage. Instead of discussing whether "using AI to see a doctor is good or bad", perhaps we should think more about "how to teach patients to use AI effectively?"

At this stage, it is still too early to say that DeepSeek can be used for diagnosis or that AI can replace professional doctors. The experience and examinations relied on in diagnosis and treatment cannot be realized by AI. At this stage, it is definitely unrealistic to say that AI can replace professional doctors, and it is far from the "sky - falling" stage. However, DeepSeek has shown the possibility of further playing a role in the medical field:

Deepening of accurate diagnosis and auxiliary decision - making: By integrating multi - dimensional data such as patients' image data, genetic information, and electronic medical records, more comprehensive diagnosis support can be provided and potential risks can be warned in time; Acceleration of drug research and development and treatment innovation: Quickly lock potential drug targets, predict efficacy and safety, and reduce the risk of failure; Popularization of personalized medical treatment and health management: Combine AI algorithms with patients' lifestyle data to generate personalized intervention plans and adjust drug dosages and treatment courses in real time; Optimization of medical resources and improvement of grass - roots capabilities: Alleviate the problem of insufficient medical resources in remote areas and assist young doctors to quickly master clinical skills; Coping with the challenges of an aging population and emerging medical needs: Through AI - driven fall - monitoring sensors, remote dialysis devices, etc., relieve the care pressure in an aging society.

At the same time, DeepSeek also proposes that when applied in the medical field, the following issues need to be noted:

Data standardization and privacy security: The fragmentation of medical data and privacy protection need to be solved through multi - party cooperation. Ethical and regulatory frameworks: Ethical norms for AI medical applications need to be formulated (such as algorithm transparency and responsibility attribution). Deep integration of technology and clinical practice: Avoid "technology islands" and promote a new model of doctors and AI working together.

Compared with offline doctor consultations, AI consultations are one - way, while offline consultations are two - way. AI can give a series of diagnosis, treatment, and medication plans. Can people choose the most suitable one? The value of this information may still need to be defined by people in the end.

DeepSeek has made a significant impact on the medical field. It has led to situations where patients question doctors' treatment plans based on DeepSeek results. While AI has advantages in some medical areas like image recognition, it also has limitations, especially in aspects like prescription writing. Multiple hospitals are accessing DeepSeek. In the medical field, AI can be applied in certain areas mainly as an auxiliary. Both doctors and patients need to adapt to the new situation of AI in medicine, and there are still many issues to be addressed regarding DeepSeek's application in the medical field.