The phrase "artificial intelligence" often brings up thoughts of advanced creations—like C-3PO, Samantha from Her, or the sinister Skynet from Terminator. Experts often laugh, or grunt in annoyance, at the apocalyptic visions put forth by Hollywood, but the laughter may one day give way to grudging respect: Computing power could soon advance so that AI is smarter than people, and connected to everything, according to Beena Ammanath, VP for data and analytics at General Electric.
GE and other industrial companies, like Siemens, are developing sensors and monitoring systems to warn of pending breakdowns in everything, from jet engines to power plants. That capability is called "narrow AI." Ammanath believes that the coming decades will bring first artificial general intelligence, which is roughly as smart as a human.
By 2040, hardware and software advances will then enable artificial super intelligence, she says. "Artificial super intelligence is as good or better than all the smartest human beings in the world," says Ammanath. This will allow machines to not only anticipate problems, but design future versions that will work better. "Can you think of a world where artificial intelligence is prevalent?" she asked. Now I can, and I just stopped laughing. SC