CassandrichPotentially uncomfortable but true: if you're doing A/B testing *of any kind* on users who did not explicitly consent to be part of such a test, you're performing unethical research on human subjects.
(Repost dug up from an old 
thread of mine.)
aes@dalias Maybe, or maybe it highlights a problem with the 'any unauthorized experiment'.
It would include things like, for example, a fruit vendor setting up in different ways to see what sells best, which has probably been done since forever and trying to ban it sounds completely deranged.
Or maybe that is unethical, but it doesn't matter, the humans are all unethical sacks of shit all the time anyway.
I don't like either of these.
So, maybe the difference is some other variable...?
cwicseolforFair, but, to play devil's advocate for a moment, what's different from getting fruit of different quality from the vendor?
Intent?
Ok, so one difference is that the fruit seller wouldn't try to sell worse product, just to see if he could get away with it.
No, wait, he would, but he'd be evil.
@aes @dalias Different quality, or maybe they’re selling me unlabeled lemonade but selling my friends unlabeled kiwi juice - nothing malicious or inferior, but say my friends are allergic, hate, or have bad memories about kiwi, and they take a nice big gulp on my ostensibly-qualified endorsement before figuring it out. In any case it’s in disregard of the customer’s ability to accurately communicate with others what they’re obtaining where, which is disruptive of interpersonal trust, but also interferes with customers’ ability to push back on bad design since they’re likely to dispute each others’ experiences. And that last really is evil.