The process was deceptively simple. The computer program randomly guessed how the genetic mechanism worked to facilitate the regeneration and then compared the estimate to a simulation of the flatworm’s genetic network. After three days, the computer came up with a plausible explanation as to how the genetic network accomplishes the regeneration.
The process was more than just number crunching, but the creation of a model that explains a natural phenomenon. Modeling is the most creative thing that scientists do, in the opinion of Levin and Lobo.
Building a Synthetic Scientist
Of course, we’re a long way from building computer scientists (i.e. computers that are scientists rather than scientists who study computers.) Levin and Lobo spent years developing the software to run the experiment, even going so far as to write their own computer language. They also had to develop a database of experiments conducted over the past 120 years that attempted to explain how a flatworm regenerates.
Nevertheless, the two Tufts scientists believe they are on to something. When a scientific problem has proven hard to crack, and the data concerning the problem is too vast for human beings to get a handle on, then computing power can be brought to bear to attack the problem quickly and efficiently. Levin is already using the approach to study how cancer spreads throughout the human body, an insight that could lead to new treatments for the deadly disease.
Can Computers Replace Scientists?
So, the question arises, are computers going to become creative enough that they can replace human scientists entirely? Some science is advanced by flashes of insight that are difficult to quantify, such as when Einstein hit upon relativity. Could computers be given the gift of intuitive thinking, combining a human trait to the impressive processing power of a machine?
Prime Mind (2) has an account of a computer program called AlphaGo, which beat world-class Go champion Lee Sodel 4 to 1 in a five-game match. AlphaGo accomplished this feat using two neural networks, a cybernetic equivalent of the human brain.
One neural network analyzed every possible move and calculated the likelihood of each of them leading to a victory. This sort of gaming approach has been used by computers for a while. The second neural network worked from a database of online games played by humans and determined the likelihood of them being successful by playing against another version of itself. The result was something approaching not just artificial intelligence, but artificial intuition.
The success of AlphaGo brings with it a terrifying prospect of people being rendered obsolete by the machines they create. Computers can now become not only scientists, such as the machine being run by Tufts, but artists, business managers, and even politicians.
The best-case scenario would involve human beings living like a slow-witted species that is tolerated and perhaps even indulged by the new cybernetic lifeforms but is no longer in charge of its own destiny. The worst-case scenario is something like the “Terminator” movies, in which the machines conclude that we humans are a pest and decide to wipe us out.
Is there a way to avoid such a dark future, short of turning Luddite and smashing all of the machines? Fortunately, such a means exists.
A recent story in Gizmodo (3) touches on an even more exotic scientific field of study called enhanced human intelligence. The idea is to use genetic modification and cybernetic implants to boost human intelligence and processing power. Such humans would be telepathically connected to the Internet (4) with instant access to all the knowledge stored within. They would combine the innate intelligence of the human brain with the processing power of computers. Such humans would be something like the Borg from “Star Trek” but retaining their individuality and moral sense.
So we will not be pets, slaves, or prey of the machines. We will, to use the well-know term from “Star Trek,” assimilate them and cause the two to become greater than they were separately.
References & Image Credits:
(5) photo credit: CINECA Building Marconi: Phase II via photopin (license)
(6) photo credit: CINECA Installation of Marconi via photopin (license)