In my work, letters to referring doctors are proof of evidence that I did something for the patient. Speech-to-text tools now turn my medical consult into words (and let me look at the actual person, which is nice). The other doctor won’t bother that it’s automatically written as long as it’s accurate. The problem for my profession will arise when a little agent swarm calculates the risks for one policy over the other, asks for a scan and prescribes medication. This will be soon I think, although the data to train those specialized swarms is still protected by law.
Given the lack of repercussions for AI companies that have hacked various services, there will be an AI company that trains their medical agents on PHI hacked from Epic/Cerner and gets away with it.
