Microsoft claims AI system better than doctors at diagnosing complex health conditions
Microsoft says AI system better than doctors at diagnosing complex health conditions
Bullshit Alarm
I don't have any actual studies to back this up (because I'm not sure they exist) but I could see this being the case, in some cases. At least in the west, Some groups who aren't white men - e.g. women or black people - are often not given accurate diagnoses by real doctors, whose training or biases cause conflicts there. If the AI diagnoses can avoid those same biases, it could actually be better for those groups especially.
I'm a white man and I promise you that it's no protection from our shitty medical system. My adult life has been a misery for at least 15 of the past 20 years because of multiple rushed and incorrect diagnoses. There are literally thousands of conditions that the system is incapable of recognizing because they involve multiple specialties, fit between specialties, or are frequently referred to the wrong specialist who won't look for causes outside their specialty.
A lot of it comes down to the limited amount of time Doctors get with the patient. They have to come up with a conclusion in 10-15 minutes, then move on, then it's usually several months before they see the patient again.
Microsoft said it was also a cheaper option than using human doctors because it was more efficient at ordering tests.
Oh for fucks sake. Might as well just add to this "news" that 11/10 insurance companies approve this message.
Zero credibility. Tech, AI, Media, America.
Meanwhile, Microsoft is forcing their employees to use Copilot because I guess not enough of them already are, while users on Github are trying to figure out if they can block it. Microsoft is not trustworthy.
Somewhat tangential, but a few years back Bedside Rounds did an episode on how some believed that a powerful enough database could replace a doctor's need to diagnose. There was some fruitful data that suggested a computer could out perform doctors, but as the project developed, further complications arose. Some thought the method was sound and the computing power was lacking. This was during the late sixties and early seventies, so the first AI wave.