A study conducted by PulseLife, a Lyon-based company specializing in digital solutions for healthcare professionals, has just published its first report on the adoption of artificial intelligence in medicine.
Conducted among 608 respondents, 54% of whom were doctors, the sample is presented as representative of the French medical population. The Yiaho team reviews these interesting figures concerning the adoption of ChatGPT in the medical field.
ChatGPT: The most popular tool despite its limitations
The striking figure: 62% of doctors report integrating ChatGPT into their daily practice.
An impressive score considering the tool has only been available to the general public for three years. More broadly, over half of the healthcare professionals surveyed admit to using a general-purpose AI (ChatGPT, Gemini, Grok, or others) in their work.
Usage is even more frequent in underserved areas: practitioners working in medical deserts use it more, for example, to compensate for a lack of time or colleagues to consult quickly.
Massive AI adoption, but limited trust
Behind this rapid adoption lies a more nuanced reality.
- 66% of users have already encountered false or incomplete answers,
- 94% say they are dissatisfied with current general-purpose AIs.
- Nearly six out of ten professionals (57%) are particularly concerned about the lack of scientific validation for the information provided.
- Only a small minority, 19%, claim to have no reservations about these tools.
Also read on this topic: There is an online medical advisor on Yiaho!
Frequency of use: AI already embedded in daily life
The study also reveals how routine AI has become:
- 81% of healthcare professionals use it at least once a week.
- 23% use it every day.
These figures indicate a profound change in work habits, even if its use remains predominantly complementary and not a substitute for human expertise.
Towards more reliable and specialized tools?
This report shows us that doctors are massively adopting general-purpose AIs while pointing out their critical flaws, particularly regarding scientific reliability. This is a strong signal for publishers already working on health-specific models, which are supposed to offer sourced and validated answers.
Several tools now exist, even in France, offering AI-assisted medical advice. For example, Doctolib recently announced AI-integrated assistance, and the startup MedGPT is trying to make a name for itself in this highly competitive field. Can ChatGPT surpass these national initiatives?
In the meantime, ChatGPT and its competitors are holding their ground… cautiously. Healthcare professionals seem to have integrated AI as a quick assistant but systematically keep their critical thinking engaged. A form of balance that could well define the future of “augmented medicine.”
Source: LeProgrès
