Emergent AI’s message about complexity and emergence seems to be something that many people want to hear. The nondeterminism of emergent systems has a special resonance in our time of widespread disaffection with instrumental reason.
The language of “society” is helping to disseminate the idea that machines might be able to think like people and that people may have always thought like machines … Because the constituent agents of emergent AI offer almost tangible objects-to-think-with, it prepares the way for the idea of mind as machine to become an acceptable part of everyday thinking.
These days, when people look at emergent computer models, they see reflected the idea that the “I” might be a bundle of neuron-like agents in communication. This sounds close enough to how people think about the brain to begin to make them feel comfortable.
An agent working for a rival corporation could be programmed to give you very good advice for a time, until your agent came to rely on him. Then, just at a critical moment, after you had delegated a lot of responsibility to your agent who was depending on the double agent, the unfriendly agent would be able to hurt you badly. So the Internet will be the site for intelligence and counterintelligence operations.
When Richard Nixon taped conversations in the Oval Office, he felt sure that only he and historians in the distant future would have access to them. I want to make sure that I can shut my agent off. I even want a periodic check on whether the agent holds any sensitive information. Sort of a security clearance for my agent.
Agents learn through receiving feedback on their performance. In some Maes programs, agents evolve … Maes believes that user trust is built up most quickly when agents can communicate the basis for their decisions … In the case of mail-sort agents, Maes says that users “gradually build up a trust relationship with agents,” much as one would with a new human personal assistant. Maes’s agents may be given a human face.
There is nothing to be afraid of. It’s not going to be HAL, but “Remains of the Day.”
Intellectual disciplines such as cognitive science and artificial intelligence, which previously supported hard mastery, are themselves undergoing changes that make it impossible for them to support a somewhat more “formalist” intellectual climate. A classical modernist vision of computer intelligence has made room for a romantic postmodern one. At this juncture, there is potential for a more welcoming environment for women, humanists, and artists in the technical culture.
Even as people have come to greater acceptance of a kinship between computers and human minds, they have also begun to pursue a new set of boundary questions about things and people. After several decades of asking, “What does it mean to think?” the question at the end of the 20th century is, “What does it mean to be alive?” We are positioned for yet another romantic reaction, this time emphasizing biology, physical embodiment, the question of whether an artifact can be a life.
We are moving toward a culture of simulation in which people are increasingly comfortable with substituting representations of reality for the real.