Computers, in the coming years, will begin to get a taste of the world, according to an IBM Corp. forecast.
They also will begin to see, hear, smell and even feel, seers at the Armonk-based company said in an annual forecast of the most significant technology trends during the next five years.
"We ask: What is the next big thing?" said Bernard Meyerson, vice president of innovation at IBM who owns more than 40 patents and was honored as the Inventor of the Year by the state Legislature in 1998. "What's going to move the world?"
The conclusion of Meyerson and others at IBM was that a revolution is brewing in "cognitive computing," an emerging field in which computers operate more like a human brain and help us think.
In the coming years, he said, computers increasingly will go beyond manipulating static data like spreadsheets and begin using sensors that provide a window into the rich world of data processed by humans through the five senses.
How will those advanced computers work?
In a telephone interview, Meyerson said the New Age computers will understand the implications of data but also be able to put them in context.
For instance, "hearing" computers may receive sounds from sensors on a hill and predict a coming mudslide, he said. Another computer, meanwhile, might be able to listen to a baby's chatter and, like a human, make a judgment on whether the child is happy or sad.
"Smelling" computers already are at work, sniffing the air in urban centers for hints of biohazards. But computers also could aid physicians in detecting the distinctive smell that warns that a person has diabetes, he said.
In theory, Meyerson said, sensors could be loaded with the distinctive chemical fingerprint of a banana, and say, the individual flavors of 50 other foods.
"After a while, you take enough likes and don't likes and they generate a profile," he said. This "machine that can optimize the joy of food" could optimize a recipe to fit the taste of an individual or balance taste with health factors.
Work on taste already is seeping out of the lab, Meyerson said, and IBM scientists have begun working with chefs at an unnamed culinary institute.
In the realm of sight, computers are learning to better recognize patterns as humans do. Possible uses include finding a face in a crowd or spotting disease in an magnetic resonance imaging scan and seeing how it fits in a patient's medical history.
Those computers of the future even could transmit a sense of touch to humans, Meyerson said.
"You're staring at pants that you want to buy online," he said. "Imagine there's a way you can mimic a texture vibrationally."
The vibrations could mimic the feel of a soft fabric or a hard fabric. Alternatively, the sense of touch could be transmitted through a set of pins.
"There are a lot of ways of attacking this issue," Meyerson said.
Compared with humans, however, the computers will be taking only baby steps at processing the massive amounts of data that humans deal with daily.
"Humans are hugely efficient in downloading data to themselves," he said.
As computers gain rudimentary senses, IBM researchers said, they also will gain increasingly fluid intellects that can help them use the information, much like IBM's Watson, the supercomputer that put questions in context to beat human contestants on TV's "Jeopardy."
"There's a reasoning component," Meyerson said of Watson. "It can rework priorities."
Still, even if Watson can beat us at "Jeopardy," the human brain remains the unchallenged model of data-processing efficiency, Meyerson said.
Although the circuits of Watson burn 80,000 watts of electricity, human brains get by with a mere 20 watts, 4,000 times less.