This is an interesting perspective, but I don’t necessarily agree with the sweeping conclusions. For example, it’s true that some types of cancer screening are controversial because they may not benefit patients. Let’s suppose that there’s a screening that truly doesn’t benefit patients. Then in an ideal world, we would not do that screening. But in a world where that screening will continue to be done, then sometimes an AI system might reduce the harm by, for example, reducing the rate of false positives, where false positives can harm patients. Thus, it’s not always true that adding AI to a non-beneficial medical procedure worsens patients’ well-being.
https://www.wired.com/story/artificial-intelligence-makes-bad-medicine-even-worse/
Christie Aschwanden
Ideas
01.10.2020 11:44 AM
Artificial Intelligence Makes Bad Medicine Even Worse
A new study out from Google seems to show the promise of AI-assisted health care. Actually, it shows the threat.
“Machine-enabled health care may bring us many benefits in the years to come, but those will be contingent on the ways in which it’s used. If doctors ask the wrong questions to begin with… then the technology will be a bust. It could even serve to amplify our earlier mistakes.”