RCR Wireless News spoke with Benjamin Samuel, a PH.D. student in the Games and Playable Media Center’s Expressive Intelligence Studio, who says his group is working to create an advanced social artificial intelligence, useful in a gaming scenario.
The technology, called CIF (Comme il faut – “As it should be” in French), works to make social interaction within games more realistic.
In order to show the new technology at work, Samuel demoed the basic game his group created, which consisted of a group of six high school students in the week before prom. The player takes control of the biggest loser of the group and is given the goal to, within seven days, have the most popular girlfriend in school and to be voted Prom King.
It’s at this point the CIF technology starts to shine. The group conducted an ethnographic study and came up with 3,500 social and cultural considerations with which to load up the players, from “been burned by a man before” to “overbearing parents” to “known sex magnet and cheater.”
Then, as the player interacts with the characters, every action is remembered and affects future actions, so in order to really succeed in the game, the player must work to keep track of all previous interactions, just like in real life!
After each interaction, there is a breakdown which shows why a character reacted in such a way, with percentage breakdowns. “20% Zach was in a good mood, 15% Zach respects Jenny, 10% Jenny is attractive,” and so on.
The group still has a long way to go, but the technology is already being used in conflict resolution scenarios and to help children with Asperberg’s Syndrome learn and play with social interaction in a safe environment.