Artificial intelligence is a sexy field. Technologists crave to make a robot that out-thinks a man.
We’ve progressed in fits-and-starts, but a few breakthroughs lately are leading some scientists to argue computers are smart enough that we can safely lay off crusty middle-managers, and replace them with machines.
“The newest space, and the one that’s most exciting, is where machines are actually in charge, but they have enough awareness to seek out people to help them when they get stuck,” he said — for example, when making “particularly complex, novel, or risky” decisions.
There have a been a string of unrelated developments lately.
Roomba, the vacuum cleaner, proved to be a big success. Stanford’s AI project has picked up, and its leaders say they want to put a robot in everyone’s home to carry out butler tasks, starting with recognizing door knobs and then opening doors. While the field has been dominated by Japan and Korea, Stanford researchers are trying to catch up, according to this NYT story. “It’s time to build an A.I. robot,” Andrew Ng, a Stanford computer scientist, told the Times. He leads the project, called Stanford Artificial Intelligence Robot, or Stair.
Meanwhile, Tellme of Mountain View supplies the voice recognition behind directory information for toll-free business listings. It now answers 74 percent of phone calls without a human’s help, up from 37 percent in 2001.
True, much of the progress seems to be coming in speech recognition, which is a limited area as AI goes. Here’s a Mercury News story about researchers at Stanford Research Institute in Menlo Park coming up with a better way for Iraqi soldiers to interview Iraqis, using an advanced translation tool. And here’s a NYT piece today talking about how speech-recognition dictation software has suddenly ready for prime time after decades of frustration. Nuance, on the East Coast, and Natspeak, of San Rafael, are two of the leaders. Finally, there’s the blind Google employee who is making Google searches easier for the blind.