Yesterday I went into lab just to hear a talk about advances in brain-machine interface research. A lot of times these talks devolve into equations and nuances of algorithms that my brain simply can not handle, but the talk yesterday was surprisingly intuitive and made me think about the future of our relationship with technology in a new way.
First off, I think the work is great.
I saw a movie of a man in his 20's who had been stabbed in the back of his neck with a knife. He was completely paralyzed from the neck down. He had opted to participate in the first phase of trials which involved implanting an array of electrodes in his brain. By recording, encoding and decoding the information from just 100 electrodes, he was able to move a cursor on the screen to check his email. Believe me, if I had been paralyzed for 3 years because someone put a knife in my spine, the first thing I would want to do is check my e-mail. No, really--I am an e-mail addict.
Would I let someone put an electrode array in my brain to do it?
. . . um, maybe.
Admittedly, it was awkward (and probably frustrating) watching the cursor float around the screen, over and under-shooting the target. It reminded me of myself playing those car-racing video games. I am a great driver, but it doesn't translate to R.C. Pro-Am. I get stuck in every single corner and end up driving backwards 75% of the time. But I imagine it had to feel good to have some way of interacting with your environment without using another person. In this way, robots are potential super-helper-buddies.
Second off, I think this work is really terrifying.
I have a particular fascination with robots, and by robots I mean any machine that performs a service for humans. Every day we use money-robots, coffee-robots, music-robots and of course, e-mail robots. I am fascinated by robots because they are incredibly powerful and efficient. They follow rules, and usually these rules are things that us clever humans have made up.
But technology is changing the rules, literally. These new robots learn to make their own rules which makes them ultimately more useful, more powerful and more terrifying. When I say terrifying, I don't mean in that Freddy Kreuger sort of way--more in that extreme vertigo, feet-half-on, feet-half-off a precipice on a windy day sort of way.
To get back to brain-machine interfaces for a moment, the reason why the whole set-up works is because the machine learns the language of the brain. This language can be messy, ambiguous and can change (neurons are fussy, they get loud, they get quiet, they disappear and they reappear without warning). The robot has to be flexible and recognize these changes and relearn what the brain is trying to say--on a very fast timescale. And this is exactly what is happening.
Needless to say, a human listening to the activity of 100 neurons would not have the slightest clue that the brain is saying "right hand corner".
So, robots are learning how we work and are learning how to do things for us without having to be told explicitly what we want. And they are doing it better than we can. This is the part that gets scary--how soon will it be before we have robots that can "telepathically" do our bidding? My mind races at all the scenarios that could be happen--not in the near future--but in the world of the theoretically possible.
Robot! Make that itch on my leg stop!
Robot! Change my facebook status to "hasn't gotten out of bed for 5 days!"
Robot! Remind me of that time I went to the beach, only this time change the memory so that instead of getting stung by a jellyfish, I grabbed on to a dolphin's fin and it whisked me away to a magical island where we did crossword puzzles and ate marzipan all day.
Okay, these examples may only be compelling to me. But just think about all the things in which a command from our brain directly leads to action, whether it's external (like typing) or internal (like remembering). Now, just take away the action part and insert a "Robot!" command. It's incredible. It's possible. It is a miracle for those who have been injured and have lost use of some part of their body, but it may present new challenges for the rest of us.
What is the solution? I have no idea. I believe that knowledge has its own inertia that can be slowed but not stopped by external forces (and external funding). We can't ignore these hypotheticals until they are real, or worse, believe that we can prevent them. Personally, I love this stuff. The fact that we are able to understand our brains at all is a gift that we should use to improve the quality of our collective lives.
I think the only danger lies in knowledge without understanding. We should probably do a little thinking about this stuff before taking the plunge. The "can we" question is always easier to answer than the "should we" question, but I think it's clear that we need answers to both--preferably at the same time.
Asimov had some very good intuition on this early on with his robot laws , but even these rules don't necessarily cover every good-robot-gone-bad scenario.
Looks like we are going to have to make up some new rules.
No comments:
Post a Comment