Robot Consciousness
February 26 2008 / by futuretalk
Category: Technology Year: 2008 Month: Feb Rating: 19
Robots could gain human consciousness by 2030, futurists predict.
By Dick Pelletier
“Ever since Bradley the household robot received his level-4
upgrade, he has been difficult to get along with. Bradley now has a
conscience and although he is still eager to keep the house
spotless, 
maintain security, and run errands; he wants more ‘alone
time’, and he recently joined a robots rights group.” The family
wonders what’s next – vacations; sick leave; going on dates?
Although this scenario may sound like fiction, it could depict a time when robots, programmed with human consciousness, will want to be treated more like us.
But before researchers can instill consciousness into robots, they must first define and understand this baffling human trait. Scientists are learning volumes about the brain – how it makes split-second decisions, how it learns from the past, and how it converts light into visual scenes. But so far, nobody can say exactly what consciousness is.
For some, deciphering the electrical pulses that travel through our brains is only half the story. The other part, more philosophical and complex, is how that brain activity translates into a person’s self-awareness and perception of the world around them.
Stanford neuroscientist Bill Newsome has spent twenty years studying how neurons encode information and how they use it to make decisions about the world. He is obsessed with understanding how consciousness arises from brain function, and feels the best way to solve that mystery is to implant an electrode into his own brain to observe how electric currents change his perception of the world.
It’s not certain that Newsome will get approval for such a radical undertaking, but if he does, his experiment won’t be in the interest of curing a disease or becoming a human machine. He’s hoping to do something broader: unravel the mystery of consciousness.
Positive-thinkers believe that Newsome’s work, along with efforts such as Microsoft co-founder Paul Allen’s project to create a map of the brain, HHMI’s Janelia Farm research to identify thoughts at the moment of creation, and the IBM-Swiss undertaking to reverse-engineer the human brain, could, by as early as 2020, answer two of the most profound questions in science: can the brain understand itself, and what is “self?”
With a better understanding of consciousness arising from these research projects, futurists Hans Moravec, Ray Kurzweil, and others believe that by late 2020s or early 2030s, technologists could build feelings and emotions into robots. In a report, The World in 2030, futurologist Ray Hammond predicts that “In the 2030s, robots will surpass the intellectual capacity of humans and become a conscious species.”
Should we fear our clever creations? Kurzweil says there’s little need for alarm. By the time robots surpass humans in brainpower, cutting-edge neural science in late 2020s and early 2030s will enable us to interface with these silicon wonders and share their intelligence. J. Storrs Hall, in his book Beyond AI, agrees that there is no risk. “By connecting our brains with tomorrow’s superintelligent robots,” Storrs says, “Humans will always stay a step ahead of their machines.”
Clearly, developing conscious robots poses unknown, possibly even dangerous consequences; but positive futurists see this as a natural evolutionary step as we edge closer into this “magical future.”
This article will appear in various print media and blogs; comments welcome. See other published work by Dick at http://www.positivefuturist.com and click on the “published work” tab.
Comment Thread (12 Responses)
-
I can just imagine robots negotiating for 2.7 minutes of reflective time per every 60 minutes worked at anything above a Level 3 cognitive task. :)
I agree that barring a cataclysmic event we’ll be able to create conscious entities and probably even embed that intelligence into robots according to the timeline that you lay out.
I tend to doubt Hammond’s prediction that robots will surpass human intelligence in 2030’s for some of the same reasons that Kurzweil tells us not to be worried (like the possibility of an AI:Human merger), but think that such a merger may only take place after a more powerful expansion of human thought due to increased networking with other existing humans and information systems. Human intelligence arises from the network of brains, info, tech and biology (MESTI)that that a human grows up in. Though he gets at Intelligence Amplification in his work, Kurzweil may be underestimating the power of the broader networked human mind. (one possible flaw in a huge, well thought out body of work)
Longer term, the likelihood of all intelligence networking faster and faster seems to be high (at least from an intuitive standpoint) and that would certainly cover Kurzweil’s proposed merger.
1) Is that the sort of magical future you’re reffering to?
2) If so, is that synonymous with the notion of a desirable singularity?
Cool post and welcome to FB!
Posted by: Alvis Brigis February 26, 2008
Vote for this comment - Recommend -
To Vis, You made some great comments and for the most part, you have pretty much described how this wild future could unfold. However, I believe that AI will outthink humans by mid-2030s or sooner. The advent of a human-level artificial intelligence – a machine that can match thought that we associate with humanity – will generate huge wealth for inventors and companies that develop it.
According to Business Communications Company, the AI market reached $21 billion last year and some forward-thinkers believe that it could double every two years over the next several decades. It doesn’t take a math wizard to see that this industry will soon generate trillions in sales; and this will quickly fast-forward the technology, which and could produce a smarter-than-human thinking robot within the time frame mentioned by Hammond in the article.
“Magical future” refers to an Arthur C. Clarke quote: “Any sufficiently advanced technology is indistinguishable from magic.” Our robotic world of tomorrow seems to fit this description.
I believe that much of the furor over the “Singularity” is just so much hoopla. J. Storrs Hall in his book “Beyond AI” says that when we reach the time that technology advances at breathtaking speeds, faster than humans with today’s minds can comprehend; we will possess the ability to enhance our neurons and interface with our super-intelligent AI creations, giving us the mind-power to easily unravel all the mysteries of an exponentially expanding technological world. No problem; no reason for concern.
Comments welcome.
Posted by: futuretalk February 26, 2008
Vote for this comment - Recommend -
I think it’s as interesting to think about the potential risks of an intelligent machine as our dependence on it. I was in Central America not that long ago, and was amazing at the knowledge the natives had about nature and survival – which plants and herbs had which medicinal values, what to eat and what not to eat, how to build shelters and boats, how to fish. I’m so dependent on modern day conveniences that I don’t know if I could start a fire in the woods without a lighter. I realize that in our society, it’s unnecessary to be a boy scout, but does that mean there’s no value to having those skills in the wilderness? What happens if AI comes to pass, and we relinquish control over all the usual tasks that we do. And then what if there’s a massive system failure? What if all the bells and whistles come to a screeching halt – I wonder if we would even know what to do at that point. We may fear it, but can we live without our technology?
Posted by: Venessa Posavec February 27, 2008
Vote for this comment - Recommend -
Most scientists do not fear that intelligent machines will develop huge intelligence levels on their own and one day will not feel they need humans and get rid of us. As this artificial general intelligence grows, it will provide the information that will enable humans to interface with their silicon cousins and share the intelligence.
However, many do fear that robot warriors and other killer ‘bots could become available to terrorists and bad guys who could use it against society. This is a real danger and should be of concern to everyone.
Posted by: futuretalk February 27, 2008
Vote for this comment - Recommend -
Newsome is my kind of guy: hands-on. I do trust neuroscientists more than most engineers and AI people on this topic. Jeff Hawkins is one person who may have enough inter-disciplinary knowledge to pull it off, on a modest scale, with his team.
All of the ambitious projects you mention are vital and necessary to understand the brain. Consciousness may be a bit tougher, or it may be easier in the end, than understanding the neuro-mechanics of brains. Understanding the brain will take many large, well-funded teams of scientists and bio-engineers. Understanding consciousness may come from the ruminations of somebody whom most of the world views as a crackpot.
Posted by: AlFin February 28, 2008
Vote for this comment - Recommend -
There is so much effort and attention towards unraveling the mysteries of consciousness, that this writer believes that it indeed could be accomplished by around 2020.
Then given a decade or so of dabbling with our understanding of consciousness and developing super-smart AI entities, I can easily believe that in the 2030s we will be able to instill consciousness into our machines.
The big questions then may go something like this: Will our silicon cousins attempt to change into a being more like us? And will we see the many advantages of our artificial creations and want to be more like them?
Kurzweil and Moravec probably have it right – by mid-century or so, we could merge completely with our artificial general intelligence and become a powerful, immortal being ready to begin scattering our populations to the stars.
Posted by: futuretalk February 28, 2008
Vote for this comment - Recommend -
if machines gain consciousness, and seem to be just like us – i wonder what the philosophical debate about the existence of a soul might be
Posted by: Zora Styrian February 28, 2008
Vote for this comment - Recommend -
Creating conscious robots could be compared with developing artificial life. Entrepreneur Craig Venter says the little-known field of “wet artificial life” is about to produce the world’s first human-made life forms.
Mark Bedau, COO of ProtoLife in Venice, Italy says “Creating artificial life has the potential to shed new light on our place in the universe. This will remove the fundamental mystery about creation and our role in the world.”
Conscious robots and artificial life forms will surely change our minds about religion and challenge our views of what it means to be human.
Posted by: futuretalk February 28, 2008
Vote for this comment - Recommend -
Just to throw this out, in a nutshell, consciousness is, IMHO, a 100 millisec delay feedback loop of everything in our consciousness (area of attention) and is reinserted into that very same area of attention after the delay. We then have as part of our awareness, the present environment from outside, the interior environment (feelings, goals) and a “new” thing which is a copy of what was going on 100 msec ago. That actually contains a degraded copy of everything going on in the last few seconds like a hall of mirrors where the speed of light is a few meters per second. We can “see” what we are doing now plus what we were doing 100 and 200 and 300 msec ago and the copy gets more rudimentary as we go further back in time. All this is unconscious activity so we experience a full picture of ourselves in a context of the last few seconds of activity, goals, thoughts, desires, etc. Without this short memory loop, we would have no information on who we are, where we are or what we are doing. Having it written down on a sheet of mental paper is not the same as having it fed back into our awareness on a continuous basis. This loop is unconscious but it is the fundamental mechanism that gives us a sense of being here. John
Posted by: john2004 February 29, 2008
Vote for this comment - Recommend -
John, your definition of consciousness seems very logical and it makes me wonder what kind of incredible computing power will be needed to analyze and duplicate this human state.
We will surely need the data-crunching abilities of quantum computing before we can make sense of human consciousness – and to create a software program that could be duplicated and installed into machines – wow! The computing power necessary to accomplish this would be mind-boggling.
Maybe we’ll need to enlist the help of our machines to generate the massive volume of intelligence needed to understand human consciousness and be able to duplicate it.
Understanding consciousness poses enormous problems, but if technologies advance exponentially like Kurzweil and other believe they will – this positive futurist believes that tomorrow’s scientists will find the solutions.
Posted by: futuretalk February 29, 2008
Vote for this comment - Recommend -
@ John & FutureTalk: I wonder if rather than cracking consciousness and remaining outside the equation, we’ll instead be engaged in a process of transferring and augmenting our consciousness. iow, will it be desirable or even possible to generate full models of consciousness without the two systems bleeding into one another? After all, it will be uber-compelling to put all of the info-tech we gradually develop to use gradually.
Posted by: Alvis Brigis February 29, 2008
Vote for this comment - Recommend -
Unraveling the mysteries of consciousness offers many possibilities.
Once human consciousness is understood (for example, to identify neuron firing patterns as they produce thoughts, emotions, and memories), this complex human trait with its myriad of activities could be simulated with tomorrow’s quantum supercomputers.
These simulations could then be digitally stored and become available for downloading into a new “housing unit” should disaster strike our human body or robot, essentially providing every person and/or machine with an indefinite lifespan – no possibility of an unwanted death ending life.
The new body’s simulated “mind” would believe that it is the same person or machine from the old body that was destroyed. Neither people nor machines would realize that they had died.
As far out as this concept seems, it may be only a matter of developing the information technologies as we trek through the decades ahead.
Posted by: futuretalk February 29, 2008
Vote for this comment - Recommend