By Sampreeti Bhattacharyya
Couple weeks ago, I was hanging out with my friend Ross Finman, an AI genius and a fellow grad student at MIT, on his birthday in some basement of a Cambridge apartment and doing our usual nerd talk. Virtual Reality (VR) is now the talk in the space. Ross is poking into the field and with all the hype going on about the Oculus Rift, Samsung Gear, Panasonic headsets and so on — we discussed the range of implications it could have. I recollected about five or six years ago while investigating into Second Life (virtual world https://en.wikipedia.org/wiki/Second_Life ), I had thought it would be really cool if one could experience it in 3D. That was before I knew much about VR. Now of course I imagine people are integrating the two and using the platform for immersive experiences which could be anything : venture through a primordial forest long gone, tour a black hole — or just chill out at the beach.
I asked Ross what he thought about it. Ross said, what would be way cool is if, unlike Second Life, your avatar could exist and operate in the virtual world even after you’ve logged off. Of course that requires the avatar has some sort of autonomy — and something more than Siri’s level of intelligence. It needs to have an adaptive learning capability that imitates you.
That “of course” triggered further thoughts in my head. Well, so the more I play, the more the avatar learns about me. Fair enough. Though, technically to really imitate me, that needs a massive amount of data from me at least in the research phase of this. Okay, let’s say that is possible. But wait… if there’s a virtual me in the cyber world, it would exist even if I die. Kind of like a cyber-spirit? So in a sense we would be immortal in that digital world-autonomous and performing despite of our death.
But what does it take to be YOU, and to be autonomous? Certainly you can’t do just what you’re told, or programmed! And a recent research at Tufts University is working on that: teaching robots to disobey humans if it’s harmful to them. But, really what if my virtual self develops enough autonomy to disobey me? Could I end up in a fight against… myself? Would we as a race battle our virtual counterparts? What if someone hacks into my avatar and learns all there is to know about me? Would the avatars hack each other?
Ross, maybe we shouldn’t let you in! JK.
This is pretty hard to picture, though. As a roboticist, I see robots fail all the time. Don’t last long without people to fix them! Traction on the algorithm side is a bit better than the hardware end, but neither is close to there yet. Imitating human characteristics and thinking that we don’t understand ourselves… no easy task! On the other hand, it is not impossible to rule it out either. Imagine how fast technology has grown in 200 years, from no light bulbs to satellites imaging every square inch of the earth. On the scale of the universe, or even human existence, that’s not long at all!
From today’s perspective, the rudimentary way of doing this (which researchers do) is a bunch of sensors to detect what excites you, what saddens you, what scares you, what relaxes you … all fed into the algorithm. Not so different from you and your best friend observing and knowing each other to small details, except the computer never forgets and never stops paying attention. It might be possible… or maybe we’ll get whole new ways of observing human response
Well, suppose we can replicate ourselves in the virtual world. In theory for now, for real later. So if a program can imitate us… aren’t we just programs? Probably all the same program with different parameters loaded. Load one set of numbers, you get me; another, you get Ross; another, you get you! How many gigabits different from each other are we? How about if we mix up some of the data… virtual genetic engineering!
But if all we are is a script, and our individuality is data fed into the script… is the virtual me any less real than the “real” me? Isn’t it my thoughts and feelings that define who I am… and the virtual me has those. Is it really me, but better? In other words we can literally create “cyber-clones” by running the multiple instances of the script .The virtual me can travel her world in microseconds. Have multiple copies and everywhere. I could be talking to you in Tokyo and hiking the Appalachian Trail and having dinner in Paris- all at the same time! Is that a more powerful, more flexible, and immortal me?
Interaction with the “real” world might be through holograms, and some robots directed by the virtual people to take care of physical repairs. But as time went on, the real world would become less and less relevant. Need to fix some solar panels now and then, maybe throw some more silicon in the hopper. But all the serious thinking would be in the virtual world. Even without virtual genetic engineering, with immortality, that world would grow faster than the real one.
But they would remain “us” in their origins, both good and bad. Would they love, and hate, just like us? Would they fight and kill, like us? What would a war be like… what would terrorism be like… when pathogens — computer viruses of today, but deadly in that world — can span the world in nanoseconds? But they won’t have to fight for resources like we do… maybe not all our characteristics will carry over.
Sounds insane! And yet… can we rule it out? We look around the galaxy, and see no evidence of life. Sure, we’d be pretty hard to find from light years away. But in a couple thousand years, for better or worse, we transformed our planet. In a few thousand more, wouldn’t we transform our solar system? Adjust the sun to meet our needs… the next step in adapting the environment to us, rather than the other way around. But there’s much older stars than ours! Why isn’t there any signs of stellar engineering? Did they die out, did they stop growing… or did they move into a realm we can’t see from afar?
Of course, I don’t want to sound negative. Just looking at Siri or Amazon’s Echo, it’s hard to feel much of the “terror”. We won’t witness anything too outlandish in our lifetimes. But, I think of where AI can go. I have heard of Hawking, Gates and Musk proposing to put boundaries on AI research. How do you enforce it?I don’t know. But I guess as roboticists, it is perhaps our duty to think about the evolution and implications of this technology- even if for the very far off future.
— — — — — — — — — — — — — — — — — — — — — —
[1] Image credit (middle)-©Oculus. Right: ©Second Life
[2] Image credit: Ross Finman
[3] Image credit: http://chriskresser.com/images/clones.jpg
