If only it was as simple as that. In medicine a wrong diagnosis can have very bad real world consequences so 90% cheaper and 99% faster has a very real possibility of being a net negative.

There are lots of ways of lightening the load with AI aid, though without fully removing the need for human eyes.

Yes, and these are typically the best systems, where the doctor is assisted rather than replaced. I've looked at several over the last years and the comparison doctor+augmentation tool to doctor alone and tool alone is invariably superior. For now I think that is the most productive and the safest route.

For breast screening, this task is high volume and low prevalence and AI can help with radiologist burnout from increased caseload.

Flipping a coin is like 99.999% cheaper and 99.998% faster than the AI, though. I mean, as long as you don't care about accuracy...

The lifetime incidence of breast cancer is 13%, and on any given annual exam it's under 1%. So you could build a >99% accurate AI radiologist by hard coding it to always produce negative results.

NeverTellMeTheOdds.exe

Errrr, except for the people who are dead or have unnecessary surgery? All of that failure is much much more expensive!

For a breast screening application, it will always be confirmed with manual review before biopsy.

90% cheaper seems unlikely given how exponentially increasing the costs of CPU power to train models are.

We're not talking about live-training of models here, we're talking about executing a pre-trained model. CPU / power needs are MUCH lower.

way cheaper than 90%, and way faster than 99%. Radiologists are crazy expensive, at least 50$ per scan. Let's say you use a 1000$ GPU (which you don't need to as these models are usually tiny in practice because of the limited size of medical datasets and because you're not doing hundreds of scans per second.. a CPU is very likely fine). Then.. just 20 scans in you make your money. Let's say the license cost is 5$ per scan. Still make your money very fast.

> Still make your money very fast.

Not much consolation for the families of the dead patients that your software misdiagnosed.

I doubt your reasoning will be much consolation to the families of people who didn't get a scan because it cost $50 rather than $5

US medical insurers are required to fully cover mammography screening at no cost to the patient.

https://www.hrsa.gov/womens-guidelines/index.html

Many americans don't have insurance.

note that GP assumed still charging $50 for some nice profit.

And didn't take into account the additional costs and risks of unnecessary treatments or surgery if the AI returns a false positive.