Does AI risk harming patient privacy or could it actually be used to improve it?
If patient information is necessary for AI, what challenges to privacy does AI present? Although it poses a risk to de-identification, from a different perspective it can also create the ability to construct new means of de-identifying data.