@ Shrutarshi Basu:

No, I don’t think I can link anything of the top of my head. But yeah, I agree that romantic love in large part is exactly that – hormones, brain chemistry and etc. But “love” as a concept is a little broader than that. I guess it depends on where you draw the line at what love is. Or whether or not you agree that “love at first sight” is actually love and not merely lust/infatuation, or if you think that love can be built over time, or that you can “will” yourself into love and etc.. I honestly don’t know. It’s all semantics really and as a CS guy I feel a bit out of my depth here. I guess this sort of this is best left to cognitive scientists and/or poets.

Do you have a reference for when you say: “But everything I have ever read on the subject suggests that love, as it is is more complex than that.” I’ve actually been getting the opposite impression — that want we call love is just combinations of brain chemistry and hormones.

@ Matt`:

Yeah, this appears to be true. I saw this blurb linked in a few other places listing this site as a source. Meh, I guess this is a clear example of “didn’t do research” on my part. This is why I’m not a journalist. :)

Also, looks like the site you linked to is of the “spoof news” variety. Just sayin’

How can we program a robot to love if we don’t fully understand this emotion ourselves?

‘Tis something that probably needs to be realised in more attempts at general AI – before we can code it, we need to make it non-mysterious (where “it” could be anything we want an AI to do).

It’s either pointless or dangerous to charge in with a half-baked notion of what’s going on inside any mental algorithm that currently looks like a black box, and equally no use trying to arrive at a working model by some method that supposedly (magically) lets us get there without full understanding.

Full-brain simulation might get there without the demystifying step, but once that’s done we’d still be shit-outta-luck for ways to improve on our own brains, and human brains are generally less expensive to make the ‘old fashioned’ way (even when leaving aside the fierce ethical implications of having a SimHuman inside a computer)