Enhancing Human-Computer Understanding
Computers have been dominating humans in chess since Deep Blue took down Garry Kasparov 3.5-2.5 in 1997. The result came as a shock at the time, but chess programs and computers have since advanced significantly while the state of human chess has declined slightly.
However, team an expert chess player with a highly advanced computer and the human-computer hybrid will dominate. Kasparov, inspired by his matches with Deep Blue, invented a new style of chess, called Advanced Chess. In Advanced Chess, the expert human determines a game plan, feeds it into the computer, which filters out blunders and suggests superior or equal moves. The final decision is left to the human.
This concept of augmented intelligence was the subject of a TED talk delivered in Auckland by Dr. Sean Gourley, CTO of Quid, a company that tries to implement and capitalize on human-computer cooperation.
Gourley contrasted his company’s technology with the Google search engine. For example, when one types in ‘NASA’ into the Google News finder, a list of 4,200 relevant and recent articles appears. However, according to Gourley, one can do better than that.
Using the Quid software, Gourley was able to first create a simple circle of dots, each of which represented a single article. From there, those dots became bigger or smaller as they related themselves to nearby dots. Then the larger dots organized themselves into a map, where some topic centers became evident, including the NASA budget, military use, the European Space Agency among others.
It generally takes about 10,000 hours of work in a single subject to become an expert. With Quid, Gourley wants to reduce the expertise time by augmenting the human intelligence with computer-generated visualizations of complex topics. While some may scoff at the idea of the 10,000-hour expertise time frame being shortened, it is worth noting that a soccer player and a programmer combined with low-performance computers beat out grandmasters/chess computers in one of the first online Advanced Chess tournaments.
Quid could be a step in figuring out the dichotomy between computers and humans. There are tasks that only computers can do, like factor large numbers in an instant and store massive quantities of data. There are things that only humans can do, like apply rules of common sense and expert intuition. Determining which task goes to which entity is usually a matter of ‘can this be programmed and iterated?’
Chess is something that can be programmed and iterated. As such, computers eventually bested humans. But when people were brought back into the equation to augment and implement the computer’s analytical intelligence, a new quality of chess is reached.
He hopes that this principle will branch out into more global areas, such as market analysis and military preparation, leaving the presentation with somewhat of an ominous statement. “You start to realize that if you can understand this with an advanced level of intelligence, perhaps you can start to manipulate it,” he said. “And if you can start to manipulate it, you can also start to realize the power of this software, for good or for bad.”
Whether it’s Quid’s software or some other human-computer co-operative facilitator, augmented intelligence could be a step in making sense the world’s complex topics.