The notion that a battle between a computer and a World Chess Champion is an equal one was effectively put to rest the day the reigning chess champion Garry Kasparov lost to Deep Blue, the chess-playing machine developed by IBM, in 1996. Kevin Warwick, professor of Cybernetics at the University of Reading, U.K., popularly known as the world’s first cyborg, believes the loss of popular interest in such battles since then confirms his belief that the gap between what the human brain can do and what computers can is rapidly shrinking.
The original cyborg
Prof. Warwick’s landmark experiments since 1998 resulted in connecting human nervous systems (his own and that of his wife, Irene Warwick, for instance) with computer networks. Since then, his work has extended to artificial intelligence, control, robotics and biomedical engineering. Is there a common thread that links these apparently disparate fields of interest? “My interest in all these fields is shaped by an abiding interest in intelligence,” says Prof. Warwick. He asks: how can we, for instance, harness artificial intelligence and robotics to compensate or do better what humans cannot do because of the “limited intelligence” of the human brain? While a large body of his initial work, which, he admits, is still in the nascent stage, is aimed at addressing people affected by ailments such as Parkinson’s disease, he is clearly looking at a much further horizon.
Rat robot
The issues Prof. Warwick is serious about are not merely about trying to “tweak the intelligence” of mortals but to get machines to do what the human brain is not capable of doing.
His work has extended to growing live rat brain cells (neurons) in a culture and then using them to drive intelligence in robots. His famous project, dubbed ‘Rat robot’, is depicted in a hugely popular video on the Internet. In the video, the rat robot looks like a fairly rudimentary robotics experiment comprising wheels and sensors. But the most striking aspect of the robot is that it is not powered by the customary micro-controllers or processors.
Instead, Prof. Warwick has implanted neurons from the brain cells of rat embryos and connected them to a 100-electrode array, which acts as the brain of the robot on wheels. The rat neurons, he explains, within minutes of being put together started linking with each other by growing what appears as tentacles. Soon, they connect with dendrites and axons, forming a mesh that attains brain-like functions in less than a week. In this experiment, the electrode array is connected to the rudimentary pair of wheels and the sensor. The sensor, when it approaches an obstruction, stimulates the electrodes, sending a message across the rat nervous system. This system, over time, learns how to not run into obstacles, in essence, using a biological brain to control an electro-mechanical device.
These ‘Rat robot’ experiments have now grown in complexity with attempts to form a neural network of 300,000 cells from the human brain. Warwick is now currently working on using human brain cells to develop a system in robots that can not only learn but also train themselves to adapt and work in the environment they live in.
Humans in cyberspace
Indeed, the path-breaking research he is doing with implants, significantly in the field of medicine, has enormously important applications in the world of healthcare and therapy. At his lecture here on Friday, Prof. Warwick took his audience through the substantive body of work he’s done since the early 2000s, when he famously demonstrated that it was possible for the brain to converse with a computer, or for the nervous system to be given an IP address and thus communicate with other nervous systems over computer networks.
Medical applications
Far from his arguably outlandish claims of how we’ll all need to become cyborgs in order to avoid being “overtaken” by technology, he showed how besides being ‘cool’ and futuristic, his work with doctors and medical researchers could solve complex health problems such as controlling or reducing tremors in patients suffering from Parkinson’s disease.
For instance, his collaborative research with doctors at John Radcliffe Hospital, Oxford, builds on existing deep-brain stimulation technologies (a medical process that involves positioning electrodes in the nervous centre of the brain) by using artificial intelligence (AI) systems.
In patients suffering from Parkinson’s, these AI systems can be taught to recognise electrical activity, and link it to the tremors, he explained. “It can learn not only to classify what a tremor is and what isn’t, but also predict the tremors, say 20 seconds ahead, and then switch on the stimulator prior to the occurrence of tremors,” Prof. Warwick explained.
Currently, the implants have to be replaced every two years, he said, because the batteries tend to run out. Given this is a difficult process that also runs medical risks including infection, Prof. Warwick and his team are trying to make the batteries last a lot longer by making the stimulator more intelligent, so it only stimulates when it needs to.
A substantial portion of his work involves self experiments, and a few done collaboratively with his wife. For instance, he demonstrates how he implanted two electrodes into his wife’s forearm and linked their nervous systems so he could feel any movement in her arms as a pulse in his brain. This simple telegraphic communication, he says is “of course, just the beginning”. “Clearly, the way ahead is to be able to communicate directly, brain to brain.”
On ethics
Prof. Warwick believes that these developments raise not only technical issues but also “deep” ethical and sociological questions. But ethical questions are not frozen in time, they depend on the social milieus in which they are raised, he observes. “What may not be socially acceptable as being good for the greater common good in one situation may be considered fine in another,” he remarked.
