While reading Carolyn Miller’s essay, “What Can Automation Tell Us About Agency?”, I was reminded of the Jeopardy tournament this past spring in which an IBM supercomputer made up 90 servers, named Watson, competed against two former human Jeopardy champions. By the end of the tournament, Watson had completely dominated the humans, besting the second place contestant by over $50,000. As impressive as Watson performed, it was by no means omniscient; during Final Jeopardy of the first match, Watson did this:
Miller notes that "if agency is a kinetic energy, it must be a property of the rhetorical event or performance itself" (147). Therefore, would Miller agree that Watson, the supercomputer, can only possess agency if it is situated in a game show environment, where it influences the responses of other contestants and the event as a whole? I would argue yes, because the computer was designed to interact in that specific environment. The agency of the computer only functions as "kinetic energy", not something possessed, but rather performed and acted out. If the computer was the only contestant on the show, it could be said that Watson holds one component of agency - the capacity to act - but ability to "instigate change (effect)" would be largely lost (146). The performance dimension of agency that Miller first focuses on can also relate to how Watson influenced the agency of the two human contestants. If Watson had answered every question correctly before the two other contestants - which would mean they were never allowed a chance to answer, or even talk - then Watson would be limiting their participation, and in effect, agency.
No comments:
Post a Comment