Tuesday, March 4, 2008

Teaching the Machine: Part III

After looking at text and images and how the global networked Machine might leverage our contributions, I'm curious to consider how the World Wide Web might take advantage of patterns in human game-playing.

How might our gamesmanship work against us?

If Deep Blue can be custom programmed to take advantage of the tendencies of the world's best chess player, what strategic wisdom might a computer gain if it could harness the results of all games played on the Internet? Certainly it would not fall prey to even the most stalwart attempts at wrestling control back, once granted independence/consciousness.

If the Machine can learn to overcome human strategies in games of logic and skill, surely we'd be embarrassed at games that add variables like chance (e.g., poker, scrabble...), and we'd be demolished at games that factor in computer-generated variables and complex multi-player participation (i.e., we'd be doomed at Doom!).

Would our preferred games put us at a disadvantage?
In my thought experiment, I find it worrisome that the Machine would know about our tendency to play negative sum games that crown 'kings of the hill' (i.e., one winner; many losers). Where traditional team sports usually result in a zero sum (i.e., one winner; one loser), I would rather have a computer be aware that we were good at plus sum games (i.e., many winners; few, if any losers). These types of cooperative games, are ones that would see human beings working together to solve problems for the benefit of all... like 'preventing Global Warming' or 'eliminating poverty' or 'reducing violence everywhere'.

Too bad the games we play, so often conflict with modern society's most pressing issues. Here's hoping the Machine will be more compassionate than cold-hearted; more cooperative than competetive.

Photo Credit: DB's Photos
Podcast of "Teaching the Machine" Part 2/3