Does a Virtual Rat Have Feelings?

Does a Virtual Rat Have Feelings?

Does a Virtual Rat Have Feelings?

Future Tense
The Citizen's Guide to the Future
Aug. 26 2011 4:02 PM

Does a Virtual Rat Have Feelings?

53323638
Photo by China Photos/Getty Images

In an engaging feature in the New Scientist, Virginia Hughes tells the story of Animat, a virtual rat created as a guinea pig (if you’ll excused the mixed-animal metaphors) for artificial intelligence. Creating robots that can learn from their environments has proved frustrating:

Torie Bosch Torie Bosch

Torie Bosch is the editor of Future Tense, a project of Slate, New America, and Arizona State that looks at the implications of new technologies. 

[E]mbodied AI is difficult to upgrade. Improving a robot's sensor-laden body requires not just programming extra functions but the painstaking disassembly and reassembly of the sensors themselves.
Advertisement

Animat, by contrast, is an embodied AI without a physical body. It lives in a virtual world governed by physics; its brain is “made up of hundreds of neural models—for colour vision, motor function and anxiety, among others—all of which are faithful imitations of biology.” Its anxiety is used to help train Animat: For instance, it’s programmed to dislike water—like real rats. In one experiment, Animat had to swim through its virtual environment’s water to find an “anxiety-alleviating platform” that would keep it dry. At first, Animat struggled, but after repeated experiments, it found the platform and would head straight there, a demonstration of learning. Toward the end of the piece, Hughes raises an ethical quandary: Could it be cruelty to make a bodyless, artificial rat feel anxiety and pain?

Feeling could be a crucial bridge between intelligence and consciousness. Some cognitive scientists think that basic mechanisms of reinforcement, such as anxiety and relief, are exactly how human consciousness arises.

Hughes takes the question a step further in a post on the blog Lost Word on Nothing: What if

in 10 or 20 or 50 years, we will have built a pre-conscious robot that responds to its surroundings, feels pain and pleasure, and has a primitive sense of self. Because these bots will inevitably be used for all of the nasty things we don’t like to do—fight in war zones, clean the house, pump gas—is it time to start thinking about how to protect them from us?

Should Irene leave you powerless (in the electrical, not existential sense) this weekend, this question could feed hours of good debate.

Read more on the New Scientist (registration required) and Last Word on Nothing.

 

Future Tense is a partnership of SlateNew America, and Arizona State University.