Artificial General Intelligence (AGI)

Continuing on a recent theme (see the blogpost Superintelligence or the new book chapter Artificial Intelligence and Pro-Social Behaviour) here is another brief rant from an email I just sent, about why I don't recommend people use the term AGI to mean "better / real AI (I'm an up-to-date futurist!)"
I wanted to explain why I dislike the term AGI.  I've found AGI to be a pejorative, a term used by people who don't respect the previous 50 years of AI research, saying that it has failed because the people doing it didn't care about the right issues.  This is completely false – first AI is succeeding hugely (see [that book chapter]) and second everything the AGI people want to do a lot of smart people have been trying to do for a long time.  Ignoring the reasons for their "failure", that is, ignoring the outcomes of their research, is unscholarly and often rooted in various types of arrogance, though sometimes in the quasi-supernaturalism which unfortunately surrounds work on concepts we deeply identify with, like ethics, consciousness & intelligence.  My sense from talking to you is that you use the term "AGI" because you've been told to, not because that's the way you think; I'm asking you to reconsider the authority of the people that have convinced you it's the right term.

PS I work on artificial consciousness and synthetic emotions.  It's not a big deal.

Comments