The big question around true AI may not be solved until the last minute, and that time may never arrive, if we don’t overcome current issues with computer security. It’s eminently popular to claim blockchain’s usefulness, but with AI, blockchain’s distributed-immutable-trust attributes may well be inevitable for its progress.
Thoughts on AI often lead to a key question. Will there be highly intelligent machines, and if so, will they replace us in everything we do, or even threaten us?
By “intelligent” machines, I don’t mean smart devices like a car navigation, or a robotic vacuum cleaner, that’s just narrow AI, only good for a particular task. I mean a conscious robot that feels and thinks creatively, which can not only combine various skills into brand new ideas, but can also be melancholy or sarcastic. A robot, which can do everything we do. The question is simple — will it exist?
Interestingly there is no consensus on the answer. The more we progress, the more burning will this question become, yet we will not know the answer until the very last minute.
The reason is that the answer does not really require much knowledge of computers or current AI algorithms, the really intriguing question is — who are we, the “target destination”?
Scientists, philosophers, and theologists have struggled to answer this for millennia, some think we are just a bunch of atoms, others think we have some “magic” ingredient, that robots will never possess.
So the answer does not depend on what you know, but what you believe. Not to get stuck in that field of “magic”, let’s get back to narrow AI, which is a reality and will be able to perform many of the tasks we now do in our jobs.
A great source of balanced views on this is a recent book by Byron Reese, a tech entrepreneur, called “The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity”. In it, Reese explains that a common mistake is to think that blue collar jobs are the most threatened. Presumably, a server who just sits in a drive through window taking orders needs to worry for his job, but a waiter who jovially suggests the wine to accompany your meal, flambes your steak, and performs the Heimlich maneuver when you get a fish bone stuck, doesn’t need to worry. It will take a long time for robots to master that, and I hope not to be dining out when they first attempt it.
We may see a geologist or an anesthesiologist being replaced by a robot way before robots take the jobs of a waiter, bus driver or a chimney sweep. Those jobs are not simple by any means, and from the perspective of computers, they are extremely complex and creative. Plus there are jobs such as ballet dancers, priests, or nannies, that we will never want machines to do. I’d never let a machine hold my baby.
So no worries if your kids won’t make it to Harvard, just suggest they look at jobs that involve a lot of human contact, are not repetitive and which require some creativity, and, some, let’s call it “human touch”. Like that waiter. The rest is indeed up to the computers. Well, there is one more big ‘’if’.
Yes, you’re reading an article on blockchain, I am getting there, just wanted to set a futuristic tone and find an excuse to clear some media nonsense out of the way. Why blockchain and AI ? You see, having a secure computing environment is a key prerequisite to any of this AI stuff, whether it’s an IoT garden sprinkler that wakes up when it’s dry, or an autonomous car, which people mistakenly think is just round the corner.
Before any of that comes to fruition, we have to trust computers, and that’s where crypto comes in very handy.. At the moment, we can’t trust them. Read the headlines — there is a major computer hack or a malfunction in the news every other day. Those stories are pretty bad, but in actual fact, it’s really nothing. Computers live in their own abstract world of zeros and ones, they compute and simulate all kinds of cool things, but can’t really touch anything in the physical world. The worst a hacker can do is turn those virtual ones to zeros. Until now, that is. If hijacked, that IoT sprinkler can get you wet, and the robotic car can get you killed.
So it’s about time to pull out cryptography, and it’s latest child, blockchain. Otherwise we may get in serious trouble, like in that Black Mirror episode where a swarm of robotic bees is made into killers by a remote hacker. None of the futuristic AI predictions that fans of the “singularity” concept fantasize about, will materialize, if we hit a wall of security issues (plus, btw, there may be other walls).
Sure, a ton of articles have been written on how blockchain and crypto are good for this or that. When you have a hammer, everything looks like a nail, and people love the blockchain hammer. But I’m claiming outright that AI will not happen without blockchain — it’s that simple and that serious.
First of all, public key cryptography will enable AI to behave steadily in expected ways, think along the lines of the Trusted Computing Group project, that has many hardware vendors like HP and Dell using encryption on hardware level, to make sure that only trusted applications run.
However there should be no “one key” to rule them all. When Tesla cars become fully autonomous and still download updates from the HQ in Palo Alto signed by that one key, I envision a hacker grabbing that key, kidnapping all the Tesla cars at once and driving them to Mordor. North Korean state-sponsored hackers may already have the key and are just waiting for that “Level 5” autonomous upgrade. Remember the Sony hack? Would this be more difficult?
Blockchain will enable swarms of robots to keep data integrity, authenticate among themselves, and not blindly trust a central server. You may have heard, about a thousand times, about the beauty of blockchain being in its decentralization. Replace beauty with inevitability here.
MIT Media Lab’s Eduardo Ferrer describes this concept from a wider angle: “The combination of blockchain with other distributed systems, such as robotic swarm systems, can provide the necessary capabilities to make robotic swarm operations more secure, autonomous, flexible and even profitable.“
So what if the general super-smart AI will one day come to life? What will it think of us, if it reads a bit of our history, learns how we treat the environment, or just listens to a commercial radio? Don’t think of it as when the Spanish conquered the Incas. Given the exponential progress of computers, it will be more like when you walk across a pathway covered in ants. You don’t give it a second thought when you tread on them
I just hope that super-AI will have a blockchain-based off-switch. If not, here is a tip for a nice last thought before we die: Remember that big question from the beginning of the article? It also works in the reverse — if that AI is born, it means that we are, like a computer, just a bunch of atoms. So we will finally learn who we are.
Miro Pikus, CTO, NKB Group
Check out my BlockDelta profile for my full contact details.