Hi ,
When is it OK for humans to intervene in tasks supposedly being performed by AI?
I posed this question to readers last week, and most were OK with the idea of human
intervention in AI tasks - or at least, OK provided the human intervention is disclosed.
Here are some of the comments I received (reproduced with permission):
I'm guessing there's a fine line on it. Like some tech products start on a spreadsheet in
the backend while developing, and then turn the whole thing into a product once they have enough customers. - Rod Aparicio
I think for me, if it was the other way round I would be more worried about the ethics. So, if I thought I was paying a person for a service, but
it was instead AI, I would feel very strongly that that was wrong. - Andrew Jones
Until our expectations normalize (e.g., obviously N was done by humans, and obviously Y was done by AI) it might be important in some cases to make clear - one way or the other - which
entity performed the labor.
For example, nobody thinks a Twinkie or a Motrin or a Camry were crafted by hand, and if it turned out they were, we might not be happy. - Jonathan Stark
It's just as wrong as if you pretended AI didn't write your book or your homework or what have you when it actually did. Like, just don't lie. Y'know? - James Turner
This leads me to my follow-up question:
Under what circumstances are you comfortable with the use of fully AI-powered technologies (without human oversight)?
Hit reply and let me know.
Talk again soon,
Dr Genevieve Hayes.
