Copy article

When AI Enters the Consultation Room, Who's Really in Charge?

ended 21. October 2025

The New York Times reports (20 October) the ‘ChatGPT for doctors’, OpenEvidence just raised $200 million at a $6 billion valuation. 

According to Daniel Nadler, one of its founders, the medical AI chatbot now supports 15 million clinical consultations monthly across 10,000+ medical centres, growing at 70,000 medical professionals monthly. Trained on prestigious medical journals like JAMA and the New England Journal of Medicine, it's free for verified medical professionals and funded entirely by advertising.

Investors are celebrating the rapid adoption. Doctors are praising the speed of access to peer-reviewed medical knowledge. 

Here’s what few are discussing: what does this mean for the human sitting in the consultation room?

When your doctor is getting real-time AI guidance during your appointment, you have no visibility into what's happening behind the screen. You don't know if the diagnosis came from your doctor's clinical judgment, the AI's pattern matching, or a combination of both. You can't ask the AI to explain its reasoning. And you certainly don't know whether the advertisers funding this "free" tool have any influence over the guidance your doctor receives.

The company claims safeguards against AI hallucination and warns about low confidence scores on its output, but medical AI doesn't just need to be accurate, it needs to preserve trust in the doctor-patient relationship, clinical autonomy, and informed consent. When 15 million consultations per month flow through a black box system funded by advertising, patients lose agency twice: once because they can't see the AI's reasoning, and again because they don't know how their information is being monetised.

The question isn't whether AI can help doctors, it's whether it makes them more brilliant at their jobs or just more dependent on systems that serve someone else's interests."

We welcome your views:

  • Should patients have the right to know when AI systems are influencing their diagnosis or treatment recommendations?
  • Is an ad-supported medical AI model fundamentally incompatible with patient-centred care?
  • How do we preserve doctor-patient trust when a third party (the AI) sits invisibly in every consultation?
  • Does faster access to medical knowledge actually improve patient outcomes, or just create the illusion of better care?

2 responses from the Newspage community

Copy all

Star Quote
Copy

Colette Mason
AI Ethics Consultant at Clever Clogs AI
Specialised training on peer-reviewed content, transparency about confidence levels, and doctors firmly in the loop making final decisions: this is what collaborative intelligence should look like.

Contrast that with the AI Wild West elsewhere: generic chatbots trained on Reddit threads, forum hearsay and random influencer blogs, and opaque systems that cannot explain their reasoning.

OpenEvidence shows you can scale AI without abandoning human judgement or professional accountability. The funding model needs scrutiny, absolutely, but the solution architecture, domain expertise plus human oversight, should be welcomed as the template, not the exception.
Copy

Editor
Kate Allen
Owner at Finest Stays
Visibility matters: both in knowing when AI is used, and in preserving the irreplaceable trust built through real, face-to-face human connection between doctor and patient. AI can be a brilliant medical ally as it never forgets research, never tires, and sharpens clinical reasoning, but it can’t replace a doctor’s gut instinct or their deep, historical understanding of each patient as a person.