I posted this as a reply three times on Alan's blogpost about machine intelligence, couldn't get the blogger to work, so gave up and posted it here...
Hi Alan – I completely disagree that robot intelligence is mostly evolved. By far the majority of both morphology and information processing is carefully designed, with only a tiny part of the parameter space left in some cases to be finalised by GAs. That's not evolution. Second, I do not see that there's any advantage to distinguishing between computation performed by natural or artificial systems, but that doesn't say anything about how their goals or their social roles as moral subjects. Did you see my blogpost this week? Is this in response to it? If not, have a look at it, at least at the slides http://joanna-bryson.blogspot.co.uk/2015/04/how-ai-is-influencing-our-future-and.html.
Comments
Re your first point - yes I concede I should not have used the word 'mostly'.
Re your second point - my blog post was written (in 2011) after a public debate in which other panellists were arguing from biological/human essentialist points of view. Questions from the audience, and my experience of public engagement in general, reveals that many people do indeed find it alarming and/or surprising that non-biological stuff can be intelligent. You and I may find it unremarkable, but we are not typical.