Reluctance to talk to PCs wastes technology

Shy: Speech-recognition software works-- if tried.

November 22, 1999|By Mark Harrington | Mark Harrington,Newsday

As voice-control companies work to hijack control of the computer away from the keyboard, one sizable barrier threatens to foil the plot: People still are not comfortable talking to technology.

Speech-recognition and voice-control technology have come a long way in just the past year, but even the industry's most ardent proponents admit to the psychological hurdles.

At the recent Speech'TEK 99 conference, much of the talk centered on why Americans remain uncomfortable talking to computers. The issue is critical in light of recently released sales figures.

Unit sales for voice-control software packages, which allow consumers to dictate documents and control some PC functions, dropped 0.3 percent so far this year to about a half-million units, according to PC Data, a Reston, Va., research company.

"The numbers suggest that it's not really a consumer mass-market item right now," said Roger Lanctot, the company's research director. "We're still in the novelty phase. We're probably a year or two away from it becoming something that's included with almost everything."

What's behind users' hesitance?

"There is a real reluctance to speak to a computer," said Bill Scholz, director of engineering at Unisys Corp.'s Natural Language Understanding division in Mal-vern, Pa. "It will dwindle over time, but only as speech applications get better."

A slowdown in buying technology caused by Y2K and the notion that new cellular and Internet technologies obviate some applications offered by voice control will ease pressure on the market in months ahead, he said.

Unisys announced a partnership with Motorola Inc. to work on Internet voice applications, which Scholz described as the market's great new hope.

Similarly, IBM announced a partnership at Speech'TEK with Motorola-rival Nokia to work on mobile speech control research and development. Key to IBM's efforts is looking beyond the luxury of voice control to functions that address physical disabilities and offer significant work efficiencies, executives said.

In a demonstration at the show of a Web browser at an IBM briefing, staffer Guido Corona demonstrated a system's ability to read the full content of a Web page. (Existing systems already allow users to navigate the Web using voice.) But such systems may need refining, as Corona demonstrated when noting that the "stop" key becomes the most important. Computerized read-back isn't yet sophisticated enough to discern key information from the less relevant, and all gets read in droning computer voices that can become annoying.

Separately, IBM showed a voice-operable soda machine that dispensed drinks on spoken cue. But don't expect to see it soon.

"There is a barrier as far as comfort levels go," said Robert Stern, senior adviser for business development at IBM's Almaden (Calif.) Research Center. "People need to adapt to it just like a tape recorder."

But, added Anne-Marie Derouault, director of strategy and alliances at IBM Speech Systems, consumers are beginning to understand the enabling power of voice control -- "you can do more, quicker. It will absolutely reach mass acceptance. If the need is there, people get accustomed very quickly."

In addition to the psychological, other obstacles to mass acceptance have included spotty translators and systems that require long training sessions to optimize recognition. Mark Skurnick, regional sales manager for Newton, Mass.-based Dragon Systems, which launched the market nine years ago, said the newest version of his company's Naturally Speaking software requires five minutes of training and has a translation accuracy rate of better than 95 percent.

Still, he noted, acceptance by the vast majority of home PC users isn't there yet, perhaps because of outdated expectations.

There is, he said, a "perception that the technology is not quite there yet. But now it's ready for prime time. People have to get used to the fact that they don't have to use the keyboard to use the PC."

Dragon expects next year to be the breakthrough year, he said.

One area in which speech recognition offers considerable promise is in toppling language barriers. Babel Technologies S.A., based in Belgium, showed a system here that can recognize speech among a range of different languages.

Vincent Fontaine, chief executive, said his company soon will install such a system in an airport near Paris to allow French- and English-speaking travelers to ask questions about flight information.

When language isn't a barrier, noise can be, and several companies at Speech'TEK showed products that isolate and purify a speaker's voice to allow computers to receive the best possible signal.

Baltimore Sun Articles
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.