A Human Hive Mind For Machines
Sep. 12th, 2006 10:53 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
I asked artificial intelligence researcher Ben Goertzel, "what do you think of the potential for artificial intelligence resulting from games such as those described in this video?" As the presenter said in that video, if there's ever a machine uprising, if their thoughts are the democratic output of our moment-to-moment voting, one thing the our machine masters might need to keep us around for would be to operate them! This is a surprisingly plausible scenario, given that a tug-of-war between humans attempting to dominate each other through the machines-- because they fear that others will exclusively do so-- might result in a six-billion-direction Mexican standoff, with the machines effectively excercising control over all humans.
I've sometimes daydreamed about the first "artificially intelligent being" paralleling the technique of Google. A robot would be remote-controlled by hundreds or thousands of people on the internet, which they watch via live webcam with additional curious onlookers. They can switch between small aspects of the robot's movement to influence at any given moment, such as its movement path, facing, or the destination of its hand in space. They "vote" in real time by attempting to control it like a video game, and its behavior expresses the average movement of their joysticks or keyboards.
It would be like this game of Pong played by a large audience. For the first time, a robot would possess decision-making, judgement, common-sense about 3D space, visual processing, and a real (albeit aggregate) personality. It would be an experiment not only in robotics, but also in sociology. I would be fascinated to discover whether groupthink would keep the robot's human hive mind from acting fickle and indecisive.
Is this anything like what AI researcher and philosopher Eliezer Yudkowski calls "collective volition"?
I've sometimes daydreamed about the first "artificially intelligent being" paralleling the technique of Google. A robot would be remote-controlled by hundreds or thousands of people on the internet, which they watch via live webcam with additional curious onlookers. They can switch between small aspects of the robot's movement to influence at any given moment, such as its movement path, facing, or the destination of its hand in space. They "vote" in real time by attempting to control it like a video game, and its behavior expresses the average movement of their joysticks or keyboards.
It would be like this game of Pong played by a large audience. For the first time, a robot would possess decision-making, judgement, common-sense about 3D space, visual processing, and a real (albeit aggregate) personality. It would be an experiment not only in robotics, but also in sociology. I would be fascinated to discover whether groupthink would keep the robot's human hive mind from acting fickle and indecisive.
Is this anything like what AI researcher and philosopher Eliezer Yudkowski calls "collective volition"?
no subject
Date: 2006-09-13 01:34 am (UTC)It would limit the damage that new, crazy, ignorant or foolish people could exact. While ensuring that good decision makers are rewarded.