Monday, September 9, 2013

Will Supercomputers Subdue the Human Race?


I refuse to accept the idea that we, humans, could soon build machines similar or, even worse, superior to us - machines that would think, understand, solve problems, have consciousness, experience pleasure and pain, feel emotions… and compete with us. In the domains of physiology, however, twenty-first-century science fiction seems to be much more hard science than bizarre fiction. Many futurologists are quite aggressive in their predictions and a few argue that computers not only will leave us behind us but they will also overpower us. Inventor Ray Kurzweil predicts that humans will be producing self-conscious robots by 2030 and that by 2046 computers we will outdo us. I hope not.
Software and hardware designers have generally programmed computer systems following standard logic, as created through time by brains that are identical to those of the designers. In other words, today's computer systems, however powerful they are, are just imitatators of our reasoning faculty, one of the top qualities of human beings.
Now some ambitious visionaries, who call themselves 'neuromorphic engineers', are targeting the design of computers, not to create machines operating as one of the qualities of the brain (that is, not as followers of the rules of logic) but as the owners of such quality ( that is, as the brain itself). Instead of imitating their properties, neuromorphic engineers hope to build devices that operate as ‘brains’.
According to Karlheinz Meier, a physicist at the University of Heidelberg and one of the leaders of the new branch of engineering, to achieve such ambitious goal the revolutionary machines should display at least three characteristics that our brain does have and current computers do not: (1) low power consumption, (2) fault tolerance, and (3) self-learning capability. While our brain consumes just twenty watts, any super -computer spends megawatts. While a faulty transistor can cripple a microprocessor, neurons repair themselves and, in selected brain regions, the nervous system may replace the dying ones. While intelligent computers must be programmed to learn, brains are self-learners by nature.
There are generous budgets and ambitious projects on neuromorphic engineering. There is going to be extraordinary progress both in robotics, in general, and in bioelectronics (the application of electronics in biological processes), in particular. Astounding developments in specialized equipment to support specific physiological functions—vision, hearing, mobility, artificial organs—will keep surprising us.
The list of Dr. Meier, however, is directional but not complete. It is obvious that our brains’ processing speed and data storage capacity cannot compete with the super-machines’ power in those areas. Life and consciousness, however, are well beyond the scope of today's science. I doubt that, between now and 2046, scientists will be able to create devices that simulate our sensitivity to pleasure and pain, or the emotions and feelings from there resulting.
It is good if this does not happen. Feelings encompass love and hatred, as well as altruism and greed. Those who watched "2001: A Space Odyssey ", the 1968 movie of film director Stanley Kubrick and novelist Arthur C. Clarke, might remember HAL, the computer that ‘plays’ in the movie, has emotions and, at some point, rebels against the ship’s crew and kills one of the travelers. In the same way as HAL wanted to take command of the spaceship, some futurists believe that robots will try to dominate the world, once they overpower us, that is, they will display, with more intensity, our hatred and greed—faults these that in modern humans seem to shade love and altruism.
HAL did not materialize in 2001 and I do not believe the 2046 super-machines will fall in love or jump joyfully when their sport team scores. (Will they have a favorite team?) Unless they are programmed to act so, they will never swear or shout if someone hits them hard enough to break a few of their integrated circuits. Consequently, I am confident that, if the computers of the future are unable to display such behaviors, much less will they want to take over the Planet.
Gustavo Estrada
Author of Inner Harmony through Mindfulness Meditation
Gustrada1@gmail.com

No comments:

Post a Comment