Kamikaze Rusher

They’re doing this regardless of what we have to say or think, so I’ll voice this:

I’m glad it’s being very restricted in scope and has some form of oversight, but I worry about the diversity in training data to cover the wide demographics and unique environment factors.

The Utah population is predominantly Caucasian but contains significant numbers of Black, Polynesian, and Hispanic persons. The model used for analyzing and scoring acne would need to be trained on these demographics. Some acne features may be easy to distinguish across most skin tones but compounding problems, such as ashy skin, could cause inaccurate scoring.

Speaking of ashy skin, Utah’s environment makes things a little complicated. The drier air and higher elevation is known to agitate sensitive skin. It’s not unusual to experience random nosebleeds if you’ve moved from humid environments. Eczema can also occur from dry, irritated skin. The AI is focusing entirely on acne and may not consider these other factors in their treatment. Medication to reduce oil production, which the body could be emitting to protect itself from the harsh environment change, could make things worse. An experienced doctor might take note of the new state resident and put them on a low dose as their body acclimates. It doesn’t seem their model will do any such thing.

I think they’ll find very mixed results and I hope nobody gets seriously hurt.