I'd like to see a "back of the envelope" analysis of how many MRI scanners would be needed to make this universal, and in particular how much helium we'd need. I'd guess it would be challenging, but the scale might be good for spurring advances in AI (edit, I had meant to say advances in MRI).
MRI is good at some things, but is often inferior to CT for other issues. You basically can't MRI the heart, but CT scanners are so fast that you can scan the heart.
A typical spine MRI that includes cervical, thoracic, and lumbar takes around 2 hours. A full body CT takes a minute or two even with older machines. So purely logistically, regular scans will be CT, with MRI only used for certain issues (they do well with soft tissue like ligament injuries).
I must have AI on the brain, I actually meant to say MRI, as in advances in scanner technology. But of course you're right it would be a very rich source of data
