The Coming Robot Evolution Race: Homo Sapiens May Have "Won" the Evolutionary Race to Perfect Humankind, but Artificial Intelligence and Robotics Will Evolve Faster and Farther. Rather Than Compete with Them, We May Do Well to Make Them Our Allies and Co-Evolve, Suggests a Technology Trend Analyst
Shaker, Steven M., The Futurist
Some people believe that humanity's evolutionary advance into the future is driven by how our genetic pool responds and adapts to climate change and cultural and societal dynamics. These external factors contributed to how we evolved in the past and became human. Extending that same evolutionary view forward by a few hundreds of millions of years, we arrive at comedic vision of our collective future: We'll have become creatures with a huge forehead for expanded cranial capacity and a small body due to lack of any manual labor, etc.
Most futurists, however, realize that we now have the means to shape and influence our own evolution and cause substantial change within periods spanning only hundreds and thousands of years. The interplay between our ability to map and manipulate our own DNA, as well as to integrate mechanical mechanisms into our own physiology, is driving this evolutionary adaptation. We will adapt our DNA to more readily accept the enhancements from nanotechnology and other bionic devices, and we'll engineer these to synch up with our DNA improvisations. As a result, humanity's evolutionary momentum will spiral quicker and quicker. Fashion, self-image, and social bonding will influence the "look and feel" as much as utility. So hopefully, humans won't resemble the Borgs of Star Trek, except for those of us making an aesthetic choice to do so.
Writers such as Joel Garreau, author of Radical Evolution (Doubleday, 2005), have suggested that accelerating technology could lead to an evolutionary bifurcation between the haves and have nots. Economic, religious, philosophical, and cultural views may prevent some geographical or demographic groups from participating in actions advancing their self-evolution.
The masses of humanity may not be able to afford such enhancements to themselves or their offspring. Those who can obtain genetic and artificial organ replacements may be able to live longer and healthier, and thus will be more likely to survive and reproduce. It is possible that, over time (that is, in much quicker periods than afforded through natural evolution), genetic differences between humans who augment and alter their genetic code may differ enough from those who do not. The variance may prevent interbreeding. This would lead to the creation of a separate new species.
Now, a new competitor is also emerging on the scene. This one is all artificial, with no flesh or DNA. The arrival and evolution of humanoid robots competing against cyborgs and those humans who have resisted change may be reminiscent of the competition between Homo sapiens, Neanderthals, Homo erectus, and the "hobbit" people of the Indonesian island of Flores.
Competition in Robotic Evolution
Homo sapiens chauvinists like to think we were the fittest for survival and outcompeted the other hominids. We did have some fine competitive traits, but our success has to do with some degree of luck.
There were two points when Homo sapiens almost went extinct. Between 195,000 and 123,000 years ago, Earth was in the middle of a glacial phase and the Homo sapiens population was estimated to have gone from about 10,000 inhabitants down to as few as 600 people. Approximately 70,000 years ago, drought may have shrunk the human population down to just 2,000 folks. However, this was soon followed by the "flight out of Africa," which led to a rapid expansion both in geography and in numbers for mankind. What a very exciting and competitive ancient world that Homo sapiens resided in! Machine evolution will be both more exciting and far more rapid.
Certainly, machinery endowed with artificial intelligence does not have to be robotic; it may be like HAL in 2001: A Space Odyssey, and reside within a computer's memory core, or be part of a networked set of computers. Robots do not need to be humanoid like the Asimo robot developed by Honda. They can be wheeled or tracked unmanned vehicles like Stanley, the self-driving car that completed the 2005 DARPA Grand Challenge race. They could have multiple legs like Boston Dynamic's famous Big Dog robot.
There are far better forms for robots than "human," depending on what the robot is designed to do. But robots that are designed to perform multiple chores previously done by humans--from throwing out the garbage to walking the dog to repairing a satellite--will likely be humanoid in nature. These humanoids would be our most immediate competitors.
Accelerating Robotic Evolution
Some scientists and science commentators have expressed skepticism that sentience could ever be created in a machine setting. They're impatient that humanistic AI has not yet been achieved, even though researchers have been aggressively pursuing artificial intelligence for decades.
Others disagree. Hans Moravec, the renowned roboticist at Carnegie Mellon University and author of Mind Children (Harvard, 1990), predicts that robots will surpass human intelligence by 2030, will develop humanlike consciousness, will be aware of the world and social interactions, and will gain the ability to replicate themselves and pace their own evolution. Physicist Michio Kaku, author of Physics of the Future (Doubleday, 2011), predicts that helpful robots performing the role of butlers and maids will be available by the year 2100. He is unsure how intelligent they will be, but they will have the capacity to mimic all sorts of human behavior.
Whether either Moravec or Kaku is off by a decade or two, or even several hundred years, it is really insignificant when compared to the glacial pace of natural evolution. In his 2000 paper "Robots, Re-Evolving Mind," Moravec compares the evolution of intelligence in the natural world with the progress occurring in the field of information technology.
Natural intelligence evolution starts from wormlike animals with a few hundred neurons occurring more than 570 million years ago. Very primitive fish that appeared 470 million years ago had about 100,000 neurons. One hundred million years later, amphibians with a few million neurons emerged from the swamps. One hundred fifty million years later, the first small mammals appeared and had brain capacities with several hundred million neurons. The bigger co-inhabitants at the time, the dinosaurs, had brains with several billion neurons.
After the extinction of the dinosaurs 65 million years ago, mammalian brains also reached sizes of several billion neurons. The first hominids of about 30 million years ago had brains of 20 billion neurons. You and I, and our contemporary human colleagues, have brains operating with approximately 100 billion neurons.
Compare this to the artificial intelligence evolutionary track beginning with the first electromechanical computers built around 1940, which had a few hundred bits of telephone relay storage. By 1955, computers had acquired 100,000 bits of rotating magnetic memory. Ten years later, computers had millions of bits of magnetic core memory. By 1975, many computer core memories had exceeded 10 million bits, and by 1985,100 million bits. By 1995, larger computer systems had reached several billion bits. By the year 2000, a few personal computer owners had configured their PCs with tens of billions of bits of RAM.
If one accepts the comparison of computer bits to neurons as described by Moravec, then the computer's growth in evolution expanded each decade what it took Mother Nature to achieve every hundred million years. Moravec calculates that human engineering of artificial intelligence is occurring at 10 million times the speed of natural evolution.
An approach to AI called embodiment, or embodied embedded cognition, maintains that intelligent behavior occurs out of the interplay among the brain, the body, and the world. Some philosophers, cognitive scientists, and AI researchers believe that the type of thinking done by the human brain is determined by certain aspects of the human body. Ideas, thoughts, concepts, and reasoning are shaped by our perceptual system--our ability to perceive, move, and interact with our world. Roboticists such as Moravec and Rodney Brooks (founder of iRobot Corp. and Heartland Robotics Inc.) maintain that, in order to achieve human-level intelligence, any AI-endowed system would have to deal with humanlike artifacts, and thus a humanoid would be the optimal robot to achieve this.
The new field of evolutionary robotics, like its namesake of evolutionary biology, relies on the Darwinian principle of the reproduction of the fittest. This view posits that autonomous robots will develop and evolve from interaction with the environment. The fittest robots will reproduce by observing their interactions with the environment and incorporating mutations that increase their survivability.
Humans will be unable to match the rapid evolutionary jumps afforded to completely artificial beings, even with advances in cybernetics and genetic engineering. Robotic humanoids will only be limited by the laws of physics and not by those of biology which even genetic engineering can't alter. Hopefully, the sort of destructive competition that eliminated the rivals to Homo sapiens in the past--including such competitors as Homo erectus and the Neanderthals--will not be repeated in the next evolutionary stage.
In the best possible future, non-altered humans, humans with cybernetic implants, and robotic humanoids will learn from each other, borrow and share technology, and engage in friendly collaboration, cooperation, and competition to benefit all. In considering which robotic designs to support or, on the national level, to fund, that seems a good ideal to aim for.
RELATED ARTICLE: The Robotic Ecosystem
The robot animal kingdom is filled with strange critters. The Boston Dynamics "Big Dog" robot (top right) can carry hundreds of pounds of equipment across hard terrain.
AeroVironment's "nano hummingbird" spy drone can fly 11 mph and hover in the face of 5 mph wind gusts.
University of Essex professor Huosheng Hu's robot fish mirrors the movement of an actual fish to an eerie degree and has been displayed at aquariums around the world.
MIT professor Sangbae Kim's cheetah robot concept would stand at 80 cm, with limbs composed of carbon fiber. The Defense Advanced Research Projects Agency recently commissioned the firm Boston Dynamics to build a cheetah robot based on a similar design.
About the Author
Steven M. Shaker is an executive in a market research and training firm. He is an authority on technology assessments, forecasting, and competitive intelligence. He is co-author, with Alan Wise, of War Without Men: Robots on the Future Battlefield (Pergamon-Brassey's, 1988) and, with Mark Gembicki, of The WarRoom Guide to Competitive Intelligence (McGraw-Hill, 1998). E-mail firstname.lastname@example.org.…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: The Coming Robot Evolution Race: Homo Sapiens May Have "Won" the Evolutionary Race to Perfect Humankind, but Artificial Intelligence and Robotics Will Evolve Faster and Farther. Rather Than Compete with Them, We May Do Well to Make Them Our Allies and Co-Evolve, Suggests a Technology Trend Analyst. Contributors: Shaker, Steven M. - Author. Magazine title: The Futurist. Volume: 45. Issue: 5 Publication date: September-October 2011. Page number: 20+. © 1999 World Future Society. COPYRIGHT 2011 Gale Group.