The Lost History of “Talking to Computers”
And What It Teaches Us About AI Exuberance
by
Book Cover & Preview Text
Many science fiction movies, TV series, and books have humans talking to computers or robots, early examples being HAL in the movie 2001 or C-3PO and R2-D2 in the Star Wars movies. Beyond the inspiration of movies and books, speech recognition was one of the earliest forms of Artificial Intelligence, which, as today, captured imaginations because it tried to emulate human abilities with computers. Today, speech recognition applications are common. Those applications include digital assistants on smartphones, transcription of medical reports, speakers that can play music or answer questions for us upon voice command, and automating telephone customer service.
The Lost History of “Talking to Computers” documents 27 years of efforts by companies and the individuals driving them to connect us to computers using human language and to make a financial success of the effort. In some years, over 300 companies believed they could profitably deliver speech recognition technology or its applications. They faced the hurdle of the cost of computing power being much higher than it is today and had to compromise in early years on their objectives, finding applications that both served a need but could be delivered at an affordable price.
This history is based on a paid-subscription, no-ads newsletter the author published monthly on commercial applications of speech recognition as his major business for 27 years starting in 1993. The book is unique in that the source was written at the time the events occurred so all quotes are current, not retrospective. The author was a journalist and conference organizer for those years and conducted many interviews with executives directly at the time. The book is thus based on 27 years of nearly full-time research. It by necessity covers only a subset of influential companies, but references an online “encyclopedia” the length of 35 books discussing in more detail all the companies covered in the newsletter each year for those who want to dive deeper than the book can cover.
Each of these companies argued they had a competitive advantage in either the technology or application design over all the other companies they were competing with. Many were able to exist for years while generating losses by convincing investors they had a valid argument. Some were valued at hundreds of millions of dollars while still losing money. Some died; some succeeded and are still active today.
The history provides a perspective on how difficult technologies evolve more slowly than we remember. with much of that history and its lessons lost. Most history of technology only documents the major successes.
The history of speech recognition and natural language understanding provides insights into today’s excitement over Artificial Intelligence. The book concludes with what the lessons of history imply about the future of talking and typing to computers and about today’s overall “AI exuberance.”
Formats
Book Details
About the Book
Much science fiction has humans talking to computers, an early example being HAL in the movie 2001 in 1968. That vision motivated many people to attempt making the imagined technology real, including the author, who founded a company developing speech recognition technology in the 1980s. Speech recognition was an early part of Artificial Intelligence, computers doing a task that had previously been exclusive to humans, and proved to be more difficult than early pioneers expected. The Lost History of “Talking to Computers” documents 27 years of those efforts; the source of the history is the author’s 309 monthly issues of a newsletter tracking the hundreds of companies trying to make a business out of this challenging technology.
Today, we talk to computers frequently on our smartphones (e.g., Apple’s Siri or Google Assistant) or to a home speaker (e.g., Amazon’s Alexa) and often when we telephone a customer service number. Most doctors use computer transcription of their notes for Electronic Healthcare Systems. Speech recognition is teaching how to speak a new language and finding specific content in a video file.
How did we get here and where will it take us? What can we learn from this history that has implications for the huge investments in more general AI today? Readers will be amazed at how many companies were inspired by this challenge beginning more than three decades ago.
About the Author
William Meisel’s background is in artificial intelligence, speech recogntion in particular. It began with a PhD dissertation on neural networks and a technical book on Computer Pattern Recognition while a professor at USC. It continued with applying those techniques as manager of the computer science division of a defense company. Dr. Meisel founded a company developing speech recognition technology and applications that lasted for two decades. He published for 27 years a newsletter on commercial applications of speech recognition that is the basis for this book. He has authored a number of books, including Evolution Continues: A Human-Computer Partnership.