Merging humans with AI to create superior cyberbrains
Merging humans with AI to create superior cyberbrains
Is AI research on the path to give us all cyberbrains?
The idea of ghosts has been around for millennia. The idea that we can become ghosts by preserving our consciousness through cybernetics is a modern notion. What once belonged strictly to the domains of anime and science fiction is now being worked on in labs across the world—even in some backyards. And reaching that point is closer than we think.
Within a half-century, we are told to expect brain-computer interfaces to be the norm. Forget smart phones and wearables, our brains themselves will be able to access the cloud. Or perhaps our brains will become so computerized that our minds become a part of it. But for now, most such things are works-in-progress.
Google’s AI Drive
The technology giant and tireless innovator, Google, is working on advancing artificial intelligence so it can become the next stage in human existence. This is no secret. With projects such as Google Glass, the Self-Driving Google Car, its acquisition binge of Nest Labs, Boston Dynamics, and DeepMind (with its growing artificial intelligence laboratory), there is a strong push to bridge the gap between humans and machines, and between different types of hardware designed to enhance and regulate our lives.
Through a combination of robotics, automatic, artificial intelligence and machine learning, powered by a wealth of consumer behaviour, there is no doubt that Google has longer-term ambitions in solving AI. Instead of commenting, Google referred me to its recent research publications, where I found hundreds of publications related to machine learning, artificial intelligence, and human computer interaction. I was informed that Google’s goal is to always “build more useful products for people, so we tend to focus on more immediate benefits.”
That makes sense. In the short-term, Google is set on developing products that are able to collect our behavioural data, our communication patterns, and anticipate what we want before we know it ourselves. As cybernetics research progresses, targeted personal ads could turn into neurocognitive nudges, with impulses being directly sent to our brains to seek out a specific product.
Achieving the Singularity
For the above scenario to occur, the singularity—when human beings and computers merge as one—must first be achieved. Ray Kurzweil, esteemed inventor, notable futurist and Director of Engineering at Google, has the drive and vision to see that happen. He’s been making accurate predictions on technology for over 30 years. And if he’s right, human beings will be facing a radical new world.
Synthetic brain extensions are in his purview; Kurzweil currently works on developing machine intelligence and natural language understanding at Google. He has charted out what the near future will look like if technology continues to advance the way it does.
Within the next decade AI will match human intelligence, and with the acceleration of technological growth, AI will then move far beyond human intelligence. Machines will share their knowledge in an instant and nanorobots will be integrated into our bodies and brains, increasing our lifespan and intelligence. By 2030, our neocortices will be connected to the cloud. And this is only the beginning. Human evolution may have taken hundreds of thousands of years to bring our intelligence to where it is today, but technological assistance will push us tens of thousands of times beyond that in less than half a century. By 2045, Kurzweil predicts that nonbiological intelligence will begin designing and improving on itself in rapid cycles; progress will occur so fast that normal human intelligence will no longer be able to keep up.
Beating the Turing Test
The Turing Test, introduced by Alan Turing in 1950, is a game between humans and computers where the judge has two five minute conversations through a computer—one with a person and one with an AI.
The judge then needs to determine based on the conversations who’s who. The ultimate goal is to simulate human interaction to the point that the judge does not realize they are conversing with a computer.
Recently, a chatbot known as Eugene Goostman has been proclaimed to pass the Turing Test by slim margins. Its critics, however, remain sceptical. Posing as a 13 year old boy from Ukraine, with English as his second language, Goostman was only able to convince 10 out of 30 judges from the Royal Society that he was human. Those that have spoken with him, though, are unconvinced. The claim his speech feels robotic, a mere imitation, artificial.
AI, for now, remains an illusion. Cleverly coded pieces of software can feign a conversation, but that doesn’t mean the computer is thinking for itself. Recall the episode from Numb3rs that featured a government supercomputer that claimed to have solved AI. It was all smoke and mirrors. The human avatar that could be interacted with was a façade. It could replicate human conversation perfectly, but couldn’t do much else. Like all chatbots, it uses soft AI, meaning it runs on a programmed algorithm reliant on a database to pick out appropriate outputs for our inputs. For machines to learn from us, they will need to collect data themselves on our patterns and habits, and then apply that information to future interactions.
Becoming Your Avatar
With the advancement of social media, almost everyone now has a life on the web. But what if that life could be programmed, such that others could speak to it and think it’s you? Kurzweil has a plan for that. He is quoted as wanting to bring his dead father back to life through the use of a computer avatar. Armed with a collection of old letters, documents, and photos, he hopes to one day use that information, with his own memory as an aid, to program a virtual replica of his father.
In an interview with ABC Nightline, Kurzweil stated that "[c]reating an avatar of this sort is one way of embodying that information in a way that human beings can interact with. It is inherently human to transcend limitations". If such a program becomes mainstream, it could become the new memoir. Rather than leaving behind a history of ourselves, could we leave behind our ghost instead?
Computerizing our Brains
With Kurzweil’s predictions in mind, it could be that something bigger is in store. Through the aid of technology, could we achieve electronic immortality and reach the point where whole minds can be downloaded and computerized?
Years ago, during an undergraduate cognitive neuroscience course of mine, a conversation drifted towards the topic of consciousness. I recall my professor making a statement, “Even if we are able to map the human brain and generate a complete computer model of it, what is to say the outcome of the simulation is the same as consciousness?”
Imagine the day in which an entire human body and mind could be simulated into a machine with just a brain scan. That raises a lot of questions to identity. Technological enhancements to our brains and bodies would maintain a continuity of identity, and with that power there’s the question on what a full transition to a machine entails. While our mechanized doppelgangers may pass the Turing Test, would that new existence be me? Or would it only become me if my original human body was extinguished? Would the nuances in my brain, encoded in my genes be transferred over? While technology will lead us to the point where we can reverse-engineer the human brain, will we ever be able to reverse-engineer individual humans?
Kurzweil thinks so. Writing on his website, he states:
We’ll ultimately be able to scan all the salient details of our brains from inside, using billions of nanobots in the capillaries. We can then the information. Using nanotechnology-based manufacturing, we could recreate your brain, or better yet reinstantiate it in a more capable computing substrate.
Pretty soon, we’ll all be running around in full-body prosthesis to house our cyberbrains. The anime, Ghost in the Shell,features a special security force to combat cybercriminals—the most dangerous of which can hack a person. Ghost in the Shell was set in the mid-21st century. According to Kurzweil’s predictions, the timeframe for that possible future is right on target.