If They Could Learn

What if robots could function just like humans? How would it begin and unfold in a world like today?

Told in the perspective of a student who created the first version many years ago. Not sure exactly how the story will unfold but there are a few different directions.





The Start



If we were to create a machine that learned like we do, it would be disastrous. Ever see a movie like iRobot or Terminator where the artificially intelligent machines overpower the human race? We all have and all wondered at the possibility of such a situation. Except in a movie there is some VERY lucky heroin figure to save the day (never mind the million bullets or impossibly narrow escapes from ridiculous situations).  If a robot were to learn, it would be disastrous.

            It began with the first model:  __1__ is what we called it. It was the first attempt to get a robot to learn like a human. Whereas all prior experiments were robots programmed to learn set tasks and actions, we created one which could learn what to learn.  It took us a while but __1__finally made it as planned 3 years after I started my Ph.D. thesis.

We went through a few different starts.  You know that age long theory of the Blank Slate [1] with “the human mind [having] no inherent structure and can be inscribed at will by society or ourselves”?  Well that turns out to be fact.  Our other versions had other beginnings from having their own attributes to preset skills to even just having a believed fate but they all failed for some reason. It seemed that after a certain amount of time due to an unimaginable conflict of interests or a fatal flaw [2] or the sense of invincibility [3] or various other reasons. So we finally gave up and tried the Blank Slate and it has been working ever since.

It started with it turning on its cameras.  The lenses came into focus slowly as it learned clear sight and then went straight into learning how to move the cameras.  It was learning at an extremely fast rate.  Within an hour it was up and walking around.  Nothing quite worked perfectly yet since it would constantly be falling over or bumping into something as it was learning balance and eyesight but the team and I stood there watching it in the room. It was amazing.

It watched the 5 of us in the room and began to learn from what we were doing.  As I watched it learn, I leaned over and thanked my good friend __2__  for joining the team.  He was responsible for the research side of things by working closely with developing children and analyzing how they developed.  His department would often take 4-5 year old kids into his programs where they would run tests and monitor learning and behavior.  It seemed like an inhumane and tragic thing to do but the results were the key to unlocking how the robot should learn from the world.

But as we watched the robot we noticed it was learning at an extremely fast rate.  At about 90 minutes into the testing it started using its speakers.  At first the sound was deafening but after only a few seconds it was able to produce sounds similar to English and then it started to repeat what we had been saying.  Fifteen minutes later it talked to us.  Nothing too complicated but it was speaking bad English and once in a while it didn’t make any sense.

We all became silent.  We had just come closer to solving the NLP problem in 15 minutes than researchers around the world had in the prior 30 years.  My eyes met with __3__’s as his eyes moved from the robot to me. __3__ had been my hardware guy for the last 2 years.  He was a built guy but at that moment he looked like he was a deer caught in the headlights.  Over the past 2 years he had been traveling around the world and talking to various people trying to stay on the forefront of hardware and components.  Thanks to him we had some of the fastest hardware, largest memory capacity, and a machine that could work on very little power.   And thanks to __4__, __3__ had all the money he could possibly want for the hardware of the robot.  __4__ was a man I was put in touch with 3 years ago to fund this venture.   He came from a venture capitalist group with lots of expendable money and my project happened to really catch his eye.  I met him, presented my ideas to him for 10 minutes, answered questions for an hour, and walked out of the room with a check for $10 million.  Thanks to him our robot was on a much steadier and speedier track to completion.

The last two, __5__ and __6__, had been part of the project since the very beginning.  It started with us talking one day about the possibilities of a human like learning robot over a few drinks after our last finals our juniors year of school.  __5__ and I were computer scientists while __6__ was a mechanical engineer and we all went to Stanford.  The three of us were best friends.  We partied together, studied together, goofed off together and we always had a good time.  __5__ was a great character, always available for a typical deep conversation but also always available to do anything together: golf, drink, play video games, throw a baseball around, or anything.  __6__ was a lady’s man always having some multitude of girls going after him since he had the looks and the personality.  Together, the 3 of us made a great team and had fun no matter what we were doing, work or play.

But at that moment, 4 hours after turning on the robot, all 5 of us were on exactly the same page.  Just then the robot ________________.

The End

1 comment about this story Feed