"Truth in bots"

The bots should announce, “I’m not a person, or if I am, I’m not allowed to act like one.”

Or, if there’s no room or time for that sentence, perhaps a simple bot at the top of the conversation. That way, we can save our human emotions for the humans who will appreciate them.

-- Truth in bots | Seth’s Blog

“If you can’t tell the difference, does it matter?”

Quacking like ducks, et cetera.

The point of the post is a bit different (it’s predicated on there being able to tell the difference – “… only a minute or two into the interaction that you realize you’re being fooled by an AI, not a caring human”), but what happens when you can’t tell the difference? Should AIs always announce themselves as AIs if they are indistinguishable from a human? Why?