> Ostensibly, because financial considerations will not be part of the input to the diagnosis routine.

Devices that overreport are likely to be more popular and more common than devices that underreport. That is in part due to the financial incentives associated with diagnosing.

Excellent counterpoint, thank you. I hate to admit to being cynical enough to be convinced.

> Devices that overreport are likely to be more popular and more common than devices that underreport.

I think you're on to something. I bet Henry Schein execs/shareholders haven't felt this much energy since patients first saw their dental issues on 36 inch TVs.

To tie in the sibling threads: devices that underreport will also be popular, just with the insurance companies after the fact to help them deny covering procedures. Your lung biopsy come back benign? CignaAI says you shouldn't have had it in the first place so out of pocket for you!

I'd like to see it change the way dental insurance works, so that e.g. an AI diagnosis will legally be fully covered by the insurance, and anything that can't be verified through that would be considered an elective procedure. In the US dental insurance works basically opposite of health insurance - they only cover the basic checkups and you have to pay for any real medical needs. It's like if car insurance only covered oil changes and tire rotations but not accidents

Isn't it already a problem now where insurance companies are denying people cover because some machine learning blackbox says so?

You want to trust a ChatGPT system to decide if you should have dental surgery? What happens when you are in crippling pain but "computer says no"?

You wont be able to ask for a second opinion on anything though. Presumably if they are all the same software they'll give the same opinion.