Kaiser has been working with Stanford for years, to train different models, especially with images. A good use-case could be for diabetic retinopathy. There a

Kaiser has been working with Stanford for years, to train different models, especially with images. A good use-case could be for diabetic retinopathy. There are more and more diabetics and this can affect vision over time, if not properly diagnosed. Kaiser has ramped up and now takes more and more eye pictures to check for diabetic retinopathy. But it is a slow and manual process, requiring optometrists and ophthalmologists to seat and careful check pictures one by one (takes about 20min per picture). It is just possible to screen as many as we should screen. Here AI could really help, and pre-scan pictures for special attention, and be able to screen order of magnitude more patients. I guess at the end it might be something like should we just screen 10 patients with a doctor, or pre-screen 10,000 in that same time, and have the doctor still see 10 patients ?