I'm a doctor, I use AI every day, but it will never replace my clinical judgment

Artificial intelligence is changing the way doctors work, from reducing paperwork to helping them stay updated with the latest medical research. Dr Nikhil Rajvanshi shares how AI has become a trusted assistant in his daily practice and why medicine will always need the human touch.

In Short

• The doctor uses AI mainly to manage research, notes and documentation
• Reduced administrative work creates more time for listening and explaining carefully
• The piece says trust, empathy and observation remain central to care

Editor’s Note: In this fortnightly column, top doctors share how they deal with their own health challenges, offering readers expert insight grounded in experience. Read more such stories here.

When people hear that doctors are using artificial intelligence (AI), the first question they often ask is, "Will AI replace doctors?" My answer is simple: No. In fact, I believe we are asking the wrong question.

The real question is: Can AI help doctors become better at what they do?

From my experience, the answer is yes.

As a paediatric pulmonologist, allergist, and sleep specialist, every day brings new challenges. I see children with asthma, allergies, chronic cough, sleep disorders, recurrent respiratory infections, and many other complex conditions. Every consultation is different because every child is different. Along with examining patients, I also spend time reviewing the latest medical research, updating treatment plans, preparing educational material for families, documenting consultations, and coordinating care with other specialists.

Medicine today is evolving at an incredible pace. New studies are published almost every day. Clinical guidelines are frequently updated, and new therapies continue to emerge. Staying current is no longer optional; it is essential. But keeping up with all of this while also seeing patients can be challenging.

This is where AI has become genuinely useful in my daily work.

I don't use AI to diagnose diseases or decide on treatments. Instead, I use it as an intelligent assistant that helps me manage information more efficiently.

For example, AI helps me summarise lengthy research papers, compare recommendations from different international guidelines, organise academic notes, prepare presentations, and draft patient education materials in language that parents can easily understand. It also helps improve documentation, making medical records clearer and more organised.

These may sound like small tasks, but together they consume a significant part of a doctor's day. When AI helps reduce that administrative burden, I gain something much more valuable: time.

• Time to listen more carefully to worried parents
• Time to explain a diagnosis without rushing
• Time to answer questions that families may hesitate to ask
• Time to reassure a child who is frightened before an examination
• For me, these conversations are just as important as prescribing the right medicine.

Parents rarely come to my clinic looking only for a prescription. They come looking for reassurance, clarity, and confidence. They want someone who will listen to their concerns, understand their child's condition, and guide them through what can often be a stressful experience.

No artificial intelligence can replace those human interactions. Medicine has never been only about science. It is also about trust.

Every patient I see has a unique story. Two children may have similar symptoms but completely different medical histories, family situations, environmental exposures, emotional needs, and treatment goals. Understanding these differences requires experience, observation, communication, and empathy.

A computer cannot notice the subtle signs that become apparent during a physical examination.

• It cannot observe how a child is breathing while walking into the consultation room
• It cannot recognise the anxiety in a parent's voice when they describe repeated hospital visits
• It cannot build a relationship with a family over months or years
• These are the moments that define good medical care, and they cannot be automated
• That is why I always remind patients that AI is a support tool, not a decision-maker
• Like every technology, AI also has limitations.

Many people assume that because AI provides quick answers, those answers are always correct. Unfortunately, that is not true. AI systems can occasionally produce inaccurate information, misunderstand medical questions, or even generate responses that sound convincing despite being incorrect.

In medicine, even small errors can have serious consequences. That is why I never accept AI-generated information without verification.

Every recommendation must be checked against trusted clinical guidelines, peer-reviewed scientific evidence, and, most importantly, the patient's individual clinical condition. The responsibility for diagnosis and treatment always rests with the treating physician, not with technology. I also believe doctors need proper training before integrating AI into clinical practice.

Knowing how to ask the right questions, evaluate AI-generated information, recognise its limitations, and identify potential errors is becoming an important professional skill.

Just as we learn to interpret laboratory reports or imaging scans, we now need to learn how to use AI responsibly. Medical education will also need to evolve.

Future doctors should understand both the capabilities and limitations of artificial intelligence. They should know when AI can improve efficiency and when it should be ignored. Most importantly, they should learn that technology should support clinical reasoning—not replace it.

Another area where I believe AI has tremendous potential is patient education.

Many parents leave the clinic with doubts that they remember only after reaching home. They search online and often find confusing or misleading information. AI can help doctors create personalised educational materials in simple language, allowing families to better understand their child's condition and treatment plan.

This can improve treatment adherence and reduce unnecessary anxiety.

AI can also support preventive healthcare by helping identify patterns, analysing large volumes of health data, and assisting doctors in recognising patients who may be at higher risk for certain diseases. These tools can improve efficiency, but they should always complement, not replace, clinical assessment.

At the same time, we must ensure that AI is used ethically.

Patient privacy, data security, transparency, and accountability should remain central to every AI application in healthcare. Patients must know how their information is being used, and hospitals should have clear policies to ensure that technology is used safely and responsibly.

Doctors also have a responsibility to be transparent about the role AI plays in their work. Technology should strengthen trust between doctors and patients, not weaken it.

Looking ahead, I believe AI will become as common in hospitals as electronic medical records or digital imaging systems. It will quietly work in the background, helping doctors organise information, reduce paperwork, improve workflow, and access medical knowledge more efficiently.

But it will never replace the qualities that truly define a good doctor.

• Compassion cannot be programmed
• Empathy cannot be automated
• Clinical intuition cannot be downloaded
• Ethical judgment cannot be generated by an algorithm
• And trust cannot be built by a machine.
• The future of healthcare is not a competition between doctors and artificial intelligence.

Instead, I see it as a partnership.

AI can process enormous amounts of information within seconds. Doctors bring experience, judgment, communication, ethics, and human understanding. Together, they have the potential to deliver safer, smarter, and more personalised healthcare than ever before.

As doctors, we should not fear technology. We should learn how to use it wisely.

Because at the end of every consultation, after the reports have been reviewed, the diagnosis has been explained, and the treatment plan has been discussed, patients do not remember the software that helped organise information.

They remember the doctor who listened, understood, reassured, and cared. That is why I use AI every day. Not because I want medicine to become less human but because I want more time to be the kind of doctor my patients deserve.

(Dr Nikhil Rajvanshi, Consultant - Pediatric Pulmonologist, Allergy and Sleep Specialist, Rainbow Children's Hospital, Delhi)

- Ends