In the Age of the Smart Machine, What are WE Good For?

by Andrew McAfee on June 1, 2010

The title of this post is the title of a talk I gave a little while back at the DC offices of Palantir Technologies. The talk grew out of a post I did for my HBR.org blog about the comparative chess-playing abilities of humans and computers, which was in turn spurred by a great article on the subject by Garry Kasparov.

Palantir invited me to come as part of their evening Palantir Night Live series to discuss human and computer strengths more generally, as well as how the two can be productively combined.

The video of the talk is now up on YouTube. It’s been broken down into seven segments. I start talking about 5:09 into the first segment.

As I tried to organize my thoughts for this event, I found them ranging far and wide over such subjects as:

All of these are covered in the talk, which was a huge amount of fun to prepare and give.

I hope you’ll take a look at the videos (1st one is below), and let us know what you think. Do you agree with my arguments and conclusions? Is there anything important that I’m getting wrong, or leaving out? Leave a comment, please, and let us know.

  • AmySenger

    Andy, you definitely painted a sobering picture for humans living in an AI/robotic world. I don't view robots as malefactors but this role seems de rigueur in the world of Hollywood and “Futurists.” Fear sells.

    There's one camp that believes that technology will propel humans into obsolescence; there's another that's confident technology will provide humans with eternal life. And then there's Hollywood who likes to make money off of Man vs. Machine battles.

    The beauty of robots is that they're not human. I always think of Joshua from War Games; robots play the game humans teach them. But I'd like to believe that robots will become smart enough to realize that blowing everyone and everything just doesn't make sense.

    However, to entertain your presupposition that humans could become slaves to robot masters (and not the reverse), is it not possible that humans could assume the equivalence of dogs to humans? In which case, that could be a pretty cush life.

  • AmySenger

    Andy, you definitely painted a sobering picture for humans living in an AI/robotic world. I don't view robots as malefactors but this role seems de rigueur in the world of Hollywood and “Futurists.” Fear sells.

    There's one camp that believes that technology will propel humans into obsolescence; there's another that's confident technology will provide humans with eternal life. And then there's Hollywood who likes to make money off of Man vs. Machine battles.

    The beauty of robots is that they're not human. I always think of Joshua from War Games; robots play the game humans teach them. But I'd like to believe that robots will become smart enough to realize that blowing up everyone and everything just doesn't make sense.

    However, to entertain your presupposition that humans could become slaves to robot masters (and not the reverse), is it not possible that humans could assume the equivalence of dogs to humans? In which case, that could be a pretty cush life.

  • Mijori23

    Someone mentioned the Magnus, Robot Fighter comics as one of our culture's representations of the man/machine future. The comic made a strong impact on my imagination as a kid too.

    Magnus was raised by robot 1A – a machine which was programmed to adhere strongly to the three laws of robotics. 1A saw how humans were becoming very dependent on machines to the point of passivity. Magnus became trained to be highly intelligent, self-reliant and physically strong enough to fight aggressive machines that intended to harm mankind. A human trained by a machine to protect humanity against machines – as a possible scenarios for the future goes, not inconceivable. And wouldn't it make a great movie? : )

    For more info about Magnus, Robot Fighter see http://en.wikipedia.org/wiki/Magnus,_Robot_Fighter. And indexes – http://darkmark6.tripod.com/magnusind.html. BTW Andrew, I very much enjoyed your 7-part talk on YouTube.

  • Mijori23

    Someone mentioned the Magnus, Robot Fighter comics as one of our culture's representations of the man/machine future. The comic made a strong impact on my imagination as a kid too.

    Magnus was raised by robot 1A – a machine which was programmed to adhere strongly to the three laws of robotics. 1A saw how humans were becoming very dependent on machines to the point of passivity. Magnus became trained to be highly intelligent, self-reliant and physically strong enough to fight aggressive machines that intended to harm mankind. A human trained by a machine to protect humanity against machines – as a possible scenarios for the future goes, not inconceivable. And wouldn't it make a great movie? : )

    For more info about Magnus, Robot Fighter see http://en.wikipedia.org/wiki/Magnus,_Robot_Fighter. And indexes – http://darkmark6.tripod.com/magnusind.html. BTW Andrew, I very much enjoyed your 7-part talk on YouTube.

  • Phil H

    I thought your presentation of some recent research was fine, but you made the same mistake that everyone makes when talking about AI – forgetting the importance of motivation and narrowness.
    Super-intelligent computers already exist and they're not scary at all. Why? Because they don't care if we turn them off. And unless you program them to care, they never will. We will continue to create computers that outperform us in ever more areas – but each computer will have its dedicated task. The doctor droid will diagnose. The judge droid will judge. At the end of the day, we'll pull the plug and go home. They will never enslave us/murder us in our sleep, because they will have no *desire* to do so – even though they may have the capacity a thousand times over.

  • Phil H

    I thought your presentation of some recent research was fine, but you made the same mistake that everyone makes when talking about AI – forgetting the importance of motivation and narrowness.
    Super-intelligent computers already exist and they're not scary at all. Why? Because they don't care if we turn them off. And unless you program them to care, they never will. We will continue to create computers that outperform us in ever more areas – but each computer will have its dedicated task. The doctor droid will diagnose. The judge droid will judge. At the end of the day, we'll pull the plug and go home. They will never enslave us/murder us in our sleep, because they will have no *desire* to do so – even though they may have the capacity a thousand times over.

  • Phil H

    I thought your presentation of some recent research was fine, but you made the same mistake that everyone makes when talking about AI – forgetting the importance of motivation and narrowness.
    Super-intelligent computers already exist and they're not scary at all. Why? Because they don't care if we turn them off. And unless you program them to care, they never will. We will continue to create computers that outperform us in ever more areas – but each computer will have its dedicated task. The doctor droid will diagnose. The judge droid will judge. At the end of the day, we'll pull the plug and go home. They will never enslave us/murder us in our sleep, because they will have no *desire* to do so – even though they may have the capacity a thousand times over.

Previous post:

Next post: