Solo computer games—at least the more sophisticated ones—are designed to present the player with the illusion that he or she is interacting with actual human (or alien) beings, and that just out of his or her sight, millions more such beings are going about their everyday lives without regard for his or her actions.
In reality, of course, the beings one meets within such games operate within tight parameters put in place by the programmers to ensure that the story line of the game remains intact. Rather than being independent agents, these computer-generated characters are more akin to actors in a play: perhaps allowed to improvise a little (depending on the director), but still bound by the script.
But scientists at five European universities have joined forces to attempt to create millions of software beings of a new kind: beings that will “live” and evolve and interact within a computer, that can reproduce and pass on their characteristics to their offspring, and that can learn and develop new characteristics as they grow.
The project is called NEW TIES, an acronym for “New and Emergent World models Through Individual, Evolutionary, and Social learning.” The project is being headed by Gusz Eiben, a professor of computer science at Vrije Universiteit in the Netherlands. The other four universities involved are the University of Surrey in the U.K., Eotvos Lorand University in Budapest, Napier University in Edinburgh and Tilburg University in the Netherlands.
NEW TIES, which has received €1.6 million in funding from the European Commission, began in September 2004 and is scheduled to run until September 2007. Its main objective is “to realize an evolving artificial society capable of exploring its virtual world and developing its own view of that world.”
To make that happen, the software beings (also called “agents”) within the computer-simulated world have been given the ability to develop a communication system or language, but not the language itself—it’s up to them to create that on their own. They don’t have names (unless, presumably, they come up with them on their own). They do, however, have gender, life expectancy, size and metabolism. They have also been given three different ways in which they can learn: individual learning, evolutionary learning, and social learning. And they have been given the ability to pass knowledge from one to the other.
The researchers will be able to study the “decision trees” that represent the minds of their creations. They’ll be able to see, for example, whether beings think sex is more important than eating, see how they learn language, watch them work together to find and store food, and determine how they figure out who’s a friend and who’s an enemy.
So far, two thousand software beings have been created in a single computer, which has been plagued by “slowness and memory leaks,” as a news story put it. (Funny, since I entered the latter half of my 40s I’ve been plagued by the same thing.)
In about a month, though, the researchers hope to scale up the project until eventually millions of the software beings are hosted in a grid or cluster of computers. Then they’ll begin to study the social interaction of their creations.
According to Eiben, although there’s been a lot of research into artificial intelligence, that research hasn’t focused on how artificial beings interact. The research could prove useful in the development of robot collectives to accomplish certain tasks, in computer games (of course) and even in fields like sociology, anthropology and political science, to simulate and thus predict and prepare for reactions to certain events. (One could envision, for example, a Saskatchewan political party running a simulation of the province in a computer and only calling an election when the software beings within are prepared to elect them.)
In their original discussions about the project, the scientists involved were even more speculative. They noted, “It would be great to see whether/when/how they start wondering about the origins of their world and themselves. Pushing it to the extreme: see if they ask themselves if WE exist?”
If this all sounds like science fiction, so it should: the scientists note that their ideas were partly inspired by the short story “Non Serviam” by the late Polish science fiction writer Stanislaw Lem.
Critics doubt anything useful at all will come of the research. But it will be fascinating to find out.
After all, we may be nothing more than software beings in God’s computer right now.
