Skip to content


January 17, 2009

So my son and I were talking the other night (he’s gotten into that teenage philosophical introspection phase; we have to keep breaking it off in the middle in order to get some sleep – he could talk for hours) and the topic of the possibility of artificial intelligence came up.

We’re pretty much agreed that a machine (that is, computational) intelligence is possible (pace Roger Penrose).

My own suspicion is that one of the things that’s really holding back the creation of a naturalistic artificial intelligence – something with its own personality and a humanoid level of complexity – is that we’re not giving our nascent AIs nearly enough sensory information, and we’re not giving them enough ability to affect their environment and experience the results. That’s how we build our own intelligences, after all – as soon as we’re born, basically, we experience an unstoppable rush of sensations through at least six channels, and when we reach out and touch something we get feedback. What we touch moves, or resists our touch, and we can tell which has happened and adjust our next move accordingly.

Even Helen Keller had smell, taste, touch, and her own proprioceptive senses – four out of six. We give our computers input through a keyboard and mouse – what is that, maybe a dozen bytes per second? A network connection is a richer source of stimuli, but the interactivity is still very weak, compared to the bandwidth of reality.

I think we should – if we’re going to engage in this project at all – build baby robots (even if the brain has to be hosted elsewhere and communicate through an umbilical), and let them interact naively with the world until they learn how to think.

Give such devices a few years of parenting, and you’ll eventually train up something that could both have a conversation with you and give a damn about the results.

2 Comments leave one →
  1. Laura permalink
    January 17, 2009 5:59 pm

    Oooo, creepy, Alan! The thought that baby bots might “give a damn”, but what if baby bots end up with the same proclivity toward a pathological lack of empathy that a certain portion of the human population has? And who would be chosen to parent these bots? How would they be parented?

    David Byrne has an interesting post on his blog from November 23rd, wondering what would happen if, through DNA developments, Neanderthals were birthed into the world today. He wonders if, but virtue of their apparent larger brain capacity, they would actually end up taking over and pushing us to extinction.

    Ah, don’t you love how science pushes us to the precipice of the abyss??

  2. January 17, 2009 10:19 pm

    Oh, I think that there’s absolutely no reason to believe that artificial intelligences would be any more sane than natural ones. They’ll run the gamut from sinner to saint, sometimes both at once – just like the rest of us.

    David Byrne’s a freakin’ genius – when he and Brian Eno get into a room together, they must issue welder’s goggles to the bystanders just to cut down the brilliance a trifle.

    Posting from the edge of the abyss…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: