>Dr. Emdad Khan                                                                                              
>Using Your Voice to Unleash the Power of the Internet
>Dr. Emdad Khan, President and CEO of InternetSpeech, founded the company
>in 1998 with the vision to develop innovative technology for accessing information
>on the Internet anytime and anywhere, using just an ordinary telephone and the
>human voice.  Dr. Khan's technical knowledge and understanding of emerging
>markets has resulted in the development of Internet Speech's first voice Internet
>product/service, netECHO, the only product available today that delivers complete
>voice Internet access.
>Dr. Khan, a frequent speaker at voice-recognition, Internet application and other
>industry trade shows and conferences, holds 14 patents, has 3 others pending
>and has published more than 40 papers on the advent of voice technology on the
>Internet, speech recognition, neural nets, fuzzy logic, intelligent systems, VLSI
>and optics.  He has taught courses at universities and industry conferences on
>VLSI, neural net and fuzzy logic. He holds a Doctorate degree in Computer
>Science, Master of Science degrees in Electrical Engineering and Engineering
>Management and a Bachelor of Science degree in Electrical Engineering.
>Dr. Khan will describe and demonstrate the technology behind InternetSpeech's
>products and how it has been received by many blind, visually impaired and
>elderly users from various organizations, including the San Francisco's Lighthouse
>for the Blind, which has incorporated the technology into their organization, the
>National Federation of the Blind, and the American Council of the Blind.

Dr. Khan began his presentation by explaining that Internet Speech is available from any telephone. This makes the Internet available to the large number of people that don't have computers, as well as the blind and those who are too far from their computers to use them. The system has modes for surfing the web, searching for information, reading email,  and e-commerce. It is one way that people can bridge the digital divide.

Then he ran a tape of a typical user logging on and using it to visit a website. The computer had a pleasant professional sounding female voice, and it took the caller only a couple of minutes to get into a mode where she was describing the page he wanted to look at. Dr. Khan explained that there was considerable AI at work deciding which words on the page to read next. Then he played a clip of a caller using the system to buy a box of everybody's favorite cookies. Again, the system handled the job without any trouble.

Dr. Khan explained that visually impaired users tend to call for about ten minutes per call, and they tend to call several times per day. People without computers tend to have calls about as long, but less of them. Highly mobile professionals tend to call for much shorter periods, and only when they can't get access to a terminal with a keyboard to accomplish their goals more conveniently. When Internet Speech was starting out, they thought their main customers would be mobile professionals. The blind marketplace was something they stumbled into after that, and it has since turned out to be huge.

Over the next year or so the company plans to expand into providing Internet boxes for blind people and support for more languages. At the moment they are growing and the future looks bright.

During Q&A the following points came up:

E-commerce via Internet Speech does take some customizing of the site to really work well. The company asks vendors to allow one to twenty-one days for doing things like getting the dialogs and menus right before going live.

Government likes the system a lot, mostly because it gives them an easy way to comply with disability access laws.

The system is currently available in English, Chinese, Japanese, and German. There are plans to expand into providing Spanish and Arabic.

Static text that is used over and over is done by recording a human reading the words. Dynamic text is done by text to speech software. The system only has one voice right now, but there are plans for more.

For more information, please visit:


Tian Harter