Dr. Kevin Warwick , PhD, a cybernetics Professor at the University of Reading, is one of the foremost authorities on cybernetic implants. He has had several computerized devices surgically implanted in him, and is planning on having magnets implanted in his hands, and more sophisticated devices implanted in his brain. In an interview with Sander Olson, Warwick explains why he believes that computers will within the next few years pass the Turing test, and why intelligent, sentient computers will soon far surpass humans.
Kevin Warwick Question: You have had two cybernetic implants. What did these implants do? The first implant was simply an RFID device. All it did was identify me to the computer in the building, which allowed the computer to open doors, turn on lights, and say hello to me. The second implant was more sophisticated, and linked my nervous system directly to a computer and the internet. As I moved my hand, we could send the neural signals to the computer. So I could actually control a robotic hand with my own hand. As I moved my hand in New York, the robot hand in England moved in near real-time, allowing me to “feel” through the hand’s fingers.
Question: Did these implants cause any problems?
There were no problems. For the second implant, the tissue grew around the implant, incorporating it into my nervous system. But both implants were temporary – the first was in for only a couple of weeks, and the second was in for over 3 months.
Question: In order to be truly useful, many of these implants will need to be permanent. Are you confident that permanent implants are feasible? Yes, we are confident that these implants could work for a year or more. Methods will need to be devised for periodically replacing implants, but that does not present any intractable problems. Creating permanent implants is not something that I have directly worked on though.Question: What is the status of Project Cyborg?
The Cyborg project is designed to explore how far we can go with these cybernetic implants. In the latest experiment, my wife and I sent signals to each others nervous systems electrically. The next step is to do the same thing, brain to brain. That is a considerably more ambitious step, but it should be doable. We are also putting magnets into fingers of some of my students, which allow them to feel the closeness of the object, or remotely feel heat. It is actually minimally invasive. So for people interested in becoming a cyborg, this is a good time to be alive.
Question: How do your implants protect themselves against biological and computer viruses?
There have been no virus problems as of yet, but that clearly is a concern. There are thousands of people who already have implants to help deal with Parkinsons disease. Potentially, a signal could be corrupted and that could cause all manner of problems.
Question: Do you have any plans to implant any more microelectronics in your body?
I had the last implant installed in 2002, so I am ready for my next implant. The 2002 implant was new at the time, but since then it has been extensively tested. There are many more people now who are interested in the possibilities, and that should push the technological development. In the short term I will have magnets implanted, but ultimately the brain implant is the one that really excites me.
Question: You first published March of the Machines in 1997. How has the AI field changed since then?
In March of the Machines, I noted that if the intellectual capacities of machines surpassed that of humans, we would have serious problems. Since then, as I predicted, AI has definitely progressed, largely in accordance with my predictions. We don’t yet have human level AI, but we soon will.
Question: Ben Goertzel has published a roadmap (http://opencog.org/roadmap/) that predicts human level AI by 2023. What is your assessment of this roadmap?
Defining “human level” is problematic because there are a variety of ways of measuring human intellectual capabilities, and in some measurements computers have clearly surpassed human intelligence. But I see the 2023 timescale as being plausible for achieving human level AI. A couple of decades after that, humans could be in a terrible situation.
Question: When is your best guess as to when computers will pass the Turing test?
We came very close in 2008, when the best machine was able to fool 25% of the humans into thinking that it was human. Next year is Turing’s 100th birthday, and we may see a computer pass the test next year. When the Turing test is first passed, I’m sure that some individuals will question the accuracy of the test. It depends largely on one’s definition of what Turing referred to as an “average interrogator”. But we are very close, and it should happen within the next couple of years.
Question: You have mentioned that a technological singularity is inevitable. What will be signs that the singularity is close?
Some claim that we can always switch an AI off, but that is specious reasoning. We cannot now switch the internet off, and this will be the case with an increasing number of technologies. We will cede more and more control to the machines, and machines will increasingly act autonomously and make their own decisions on matters. I think that these signs that the singularity are close are going to be fairly obvious to the AI intelligentsia, but not necessarily to the general public.
Question: What is the best path towards AGI? Do you favor brain emulation or other approaches?
I don’t believe that emulating, copying, or simulating the brain will be a fruitful approach in the short run – the brain is simply too complicated. Artificial general intelligence, different from anything seen in animals or humans, will emerge on a network of machines. I can see the first AGI emerging on some server farm running some form of neural net software.
Question: What do you think will be the most important tech innovation during the next decade?
Efficient, long-range wireless power transmission will be developed within the next decade. This is a technology that Nikola Tesla began researching a century ago, and it could utterly transform our transportation infrastructure. Imagine having remotely powered cars, helicopters, and planes that are autonomously controlled by AIs. Reliable and efficient long range wireless power transmission will profoundly change the world within the next decade.