From R2D2 to The Terminator, and everything in between – Sci Fi fans around the world have fallen in love with artificial intelligence. Ready to have your mind blown? We’re not too far off from making the AI in Sci Fi, a reality. Unfortunately like with most things that could help humanity, AI research is very poorly funded.
Recently, I was able to get in touch with Ben Goertzel. Ben started the OpenCog project (an open sourced AI non-profit), acts as an adviser to the Singularity University, and currently bounces back between Hong Kong and Maryland building in-game AI. His full bio can be seen here.
Hi Ben, thanks for joining us today.
Thanks Jason — it’s an honor and a privilege ;)
1) You’ve been in the industry for quite awhile now. Did you always know that you wanted to work with AI? Was there something in your life that triggered it?
I grew up on SF, and I always knew I wanted to spend my life making SF stuff become real. But at the start I wasn’t sure if my focus would be AI, time machines, quantum gravity computers, interstellar spacecraft, genetically engineered creatures, psi powers, or what. I chose to focus on AI — initially in my late teens, and then with greater emphasis starting in my late 20s — because it seemed like the logistically easiest of the bunch: if I’m right, you don’t need any special hardware, just the right computer code and a big server farm of commodity PCs. And the more I thought about the problem, the more I felt like I might really know how to make it happen…
2) If you were talking to 10 year olds, how would you sum up the Technological Singularity in a way they’d understand?
Within a few decades, we’re probably going to have computers and robots way smarter than people in every way, inventing new stuff constantly and transforming the world in ways mere humans can’t even imagine. Hopefully we’ll be able to upgrade our brains or turn ourselves into robots and become super-smart, super-powerful beings ourselves. We don’t have this technology now, but it’s coming pretty soon, because new inventions keep getting made faster and faster — because the more inventions you have, the faster you can make new inventions, since the old inventions are tools you can use to help make new ones.
3) I know that you currently work on video game AI which is intriguing to quite a few of us. In Quake Live you can currently set up bots that run around the map and shoot at you. You can also group up with pets and NPCs in MMORPGs who help you attack bad guys on missions. But outside of that, they’re pretty dumb. Will in-game bots ever be indistinguishable from humans?
In-game bots will eventually be way smarter than humans. But right now what I’m working on is an in-game bot that’s more like a little in-game kid. So it can do smart stuff in the game — but can also learn from the human game players. The game world can be a tool for helping the AI to get smarter, and all the game players will be not only having fun, but also teaching the AI as they play the game…
4) What would a robot or program for that matter need to do in order for you to kick your feet up and say “I finally did it?” Teach itself how to tie a pair of shoes? Get your Dry cleaning? Pick up your kids from school? Explain string theory while building a fusion reactor?
World peace, immortality for all who want it, everybody liberated from needing to work for a living, starships traversing the galaxy, and brain implants or robot bodies for everyone who wants to become a superhuman AI. Thing is, I think all that has a decent chance of coming within a couple decades after we create the first AI that’s as smart as, say, an average science professor. And I think it won’t be that big a leap from an AI that’s like a human toddler, till we get that AI science professor. The big leap, I think, is from where we are now, to having an AI that’s like a human toddler. When someone gets an AI at that level, I’ll be incredibly excited — and a bit scared.
5) Will we see the advanced R2D2 and C3P0 type of AI in our lifetime?
Almost surely… and better. Not sure about the six million forms of communication though — there may not be that many useful alien races around. And the real-life R2D2 might go to the Radio Shack and buy himself a speech synthesizer…. But in concept — yeah.
6) A lot of people want to get involved with AI, and futurology, but not everyone is a scientist or a programmer. What can the average person do to help the cause? If someone reads this and does happen to have deep pockets, and wants to put their loot to good use, where would you direct them?
The OpenCog Foundation — the AI project I co-founded, aimed at creating beneficial AGI at the human level and beyond — is taking donations. We have basically no overhead, all donations go toward hiring AI programmers to work on helping build the OpenCog thinking machine. And we’d like to go in a robotics direction next year, but for that we need to buy some robots — I’ve been talking to David Hanson a lot, and would like to get some Hanson Robokind humanoid bots up and running with OpenCog.
Apart from donating money, we can definitely use anyone who can program C++ or Python — and has undergrad-level knowledge about algorithms and data structures — to help with the OpenCog project on a practical level. Apart from those possibilities, if you want to help out, I guess you’ll have to be creative — what we honestly need most is more programmers, and more money to hire said programmers. We know what design we need to implement to make a thinking machine, and hardware is not that expensive these days, so it’s not our bottleneck at this point. But of course, we can always use folks to spread the word about AGI, by writing articles, giving talks or just yakking to their friends…
6b) In an interview you did on Singularity 101 you said that you had thought a lot about how you would spend 100 billion dollars. Outside of buying a private island, how would you spend it?
If I had 100 billion dollars, I’d start a semi-private city, devoted specifically to beneficial development of advanced technologies, with residency offered to folks displaying qualification and passion for said technology development.
If not that, I’d start a massive research grant funding project for AGI, life extension, nanotech, femtotech, mind uploading, and so forth. If these areas got attention in the way that cancer research and semiconductor manufacturing do, progress would be way further along.
7) Outside of funding, what is the current biggest hurdle in your field?
Finding people who are really good at programming and also have a deep understanding of cognitive science and AGI theory, is always a challenge.
Conceptually, I think the biggest challenge facing the field is the integration of algorithms for abstract cognition, with algorithms for low-level perception and action. Right now we have good approaches for all of these, but nobody has made them all work together in a unified way. I think the OpenCog architecture solves that problem, but we haven’t definitively demonstrated that yet.
8) Since we’re a Sci Fi site, I have to ask. What do you say to the people who feel this will turn into a colossal fuck up like Cyberdyne’s Skynet division?
Setting aside the “naked guys traveling back in time” aspect … I think the possibility of advanced AIs running amok and killing everyone is a real one, it can’t be ruled out. But I think we can act to minimize the probability of this happening, by building our AI systems with rational minds and benevolence-oriented goal systems, and raising our young AIs with kindness and teaching them well. Like any other advanced technology, the potential benefits are huge and so are the risks. But that’s how humanity has been rolling for a long time — nobody could stop the development of advanced AI and robotics now even if they wanted to, at least not without destroying all of civilization. We’re on the verge of the next step in the evolution of intelligence on Earth, so we may as well embrace and enjoy it, and try to nudge it in a positive direction insofar as we can.
Thanks for taking the time Ben! I’d like to close by adding it’s become a goal of mine to start getting people excited about the future. What we as humans have the ability to achieve is astounding. But in order to make moon bases, ship people to Mars, and develop a cure for cancer, we need to get excited about science and technology. Sadly, the people who usually care about this stuff are less vocal than the rest. Imagine what would happen if people tweeted about space and tech more than they shared lolcats? When Neil Degrasse Tyson pleaded to get an increased space budget, his words were heard by less than 1/20th of the people who watched Justin Bieber’s latest video on Youtube. This shouldn’t be happening.
I’m not a programmer, an engineer, or a scientist, but just like you I have an internet connection. And I know that every time I upvote something about the future on reddit, Digg a scientific break through, or simply hit the “like” button on Facebook – I’m potentially reaching one more person. And that “one” person might just be the one who develops a warp drive. If we can make it a collective effort to get the word out, there is a good chance we’ll see Sci Fi become a reality much much sooner.
If you have any other link recommendations for this article, feel free to drop them in the comments below.