[wcoenen]

> For us to consider a machine sentient, it needs to function in a similar way to us, or else our definition of sentience gets way too broad to be true.

Imagine a more technologically advanced alien civilization visiting us. And they notice that our minds don't function quite in the same way as theirs. (E.g. they have a hive mentality. Or they have a less centralized brain system like an octopus. Or whatever.)

What if they concluded "Oh, these beings don't function like us. They do some cool tricks, but obviously they can't be sentient". I hope you see the problem here.

We're going to need a much more precise criterium here than "function in a similar way".

[cmdli]

I mean, if we found a planet that was filled with plant life, that doesn't seem to display any level of thought, speech, or emotion, would we consider that sentient? Do we consider trees to be sentient? On some level, self-similarity is the only metric that we have; other things could be sentient but we have very little way of knowing.

My belief is that we need to see similar functionality to be sure of sentience. This may exclude some things that may theoretically be sentient, but I don't think we have a better metric than functionality that doesn't also include a lot of definitely-not-sentient things.

[superb-owl]

While similarity to humans is a good indicator of consciousness, it's very much not a necessary condition!

We need to err on the side of sentience if we're going to avoid moral catastrophe.

So yes, I do take the possibility of sentient trees pretty seriously. They're much more alive than we realize--e.g. they communicate with one another and coordinate actions via chemical signals. Do they feel pain and pleasure? Who knows. But I'm definitely not going to start peeling the bark off a tree for shits and giggles.