A comprehensive analysis of 15 studies has revealed that patients rate AI-powered chatbots, such as ChatGPT, as warmer and more empathetic than human doctors in text-based interactions, raising important questions about the future of communication in healthcare.
Article contents:
- Empathy: a human skill where artificial intelligence excels?
- Study findings: AI excels in text-based empathy
- Areas where artificial intelligence shone
- Why does artificial intelligence seem more empathetic?
- Study limitations: text versus full human interaction
- The practical path: collaboration between doctor and AI
- Frequently asked questions
Empathy: a human skill where artificial intelligence excels?
Empathy has long been considered an essential human skill that artificial intelligence cannot replicate, serving as a crucial element in the doctor-patient relationship. However, a new large-scale study challenges this assumption. A comprehensive analysis led by researchers at the Universities of Nottingham and Leicester in the United Kingdom found that artificial intelligence, at least in text-based interactions, can be perceived as more empathetic than human healthcare providers. This discovery opens the door to a debate about the role of technology in the human aspects of medicine and challenges previous assertions that empathy is an exclusively human skill.
Study findings: AI excels in text-based empathy
The meta-analysis included 15 studies comparing AI responses with those of doctors and nurses to patient queries. In the largest dataset comprising 2,164 live patient interactions, and across multiple smaller studies, tools like ChatGPT outperformed human doctors by about two points on 10-point empathy scales. Overall, AI had a 73% chance of being judged more empathetic in direct comparisons. The researchers noted that “in text-only scenarios, AI chatbots are frequently perceived as more empathetic than human healthcare providers.”
Areas where artificial intelligence shone
Across nine separate studies covering cancer care, thyroid disorders, mental health, autism, and general medical questions, the ChatGPT-4 model routinely outperformed licensed physicians. For example, on thyroid surgery questions, AI empathy ratings were significantly higher than those of human surgeons. When responding to patient complaints directed through hospital departments, the gap widened significantly in favor of AI compared to patient relations staff. Interestingly, even doctors themselves rated the AI’s tone as more empathetic than their peers’ responses to the same questions in some studies, although there were exceptions in certain specialties like dermatology.
Why does artificial intelligence seem more empathetic?
Part of the AI’s advantage is structural. Large language models are trained on vast amounts of human conversation and can default to patient-centric language: validating emotions, summarizing concerns, and providing clear next steps. Importantly, AI does not feel stressed, exhausted, or defensive, and it can consistently apply the best wording. In contrast, doctors responding to messages juggle time pressure, medico-legal caution, and the complex context of a patient file, which can lead to terse or technical responses lacking human warmth.
Study limitations: text versus full human interaction
It is essential to note that every study in the review evaluated text-based interactions only. In real clinics, a doctor’s gesture, eye contact, or silence can convey care as powerfully as phrasing. Therefore, these results tell us about “written bedside manner,” not the full fabric of human presence. What was measured was perceived empathy rather than actual clinical impact. Additionally, an empathetic delivery does not mean much if the facts are wrong. AI hallucinations (fabricating information) or omissions can vaporize any good intentions, so medical accuracy remains paramount.
The practical path: collaboration between doctor and AI
Rather than replacing doctors, the authors suggest a collaborative workflow: doctors write the core medical advice while the AI refines the tone, adds clear and reassuring phrasing, and anticipates common concerns. Doctors then review this data and send it to patients. This approach can alleviate inbox burden, reduce terse responses that damage relationships, and raise patient satisfaction without sacrificing accuracy. Voice is the next frontier; if the AI’s empathy feature survives the leap from text to speech, the implications for virtual consultations could be massive.
Frequently asked questions
Q: Does this mean AI is better than doctors?
A: No, it means AI is better at crafting text responses that appear more empathetic and warmer to patients compared to responses often written by busy doctors.
Q: Why does AI excel in text-based empathy?
A: Because it is trained to consistently use patient-centric language and is not affected by the fatigue or time pressure faced by human doctors.
Q: Can these findings be applied to face-to-face interactions?
A: No, the study focused only on text-based interactions. Face-to-face interactions involve body language and eye contact, aspects not evaluated in this study.
