Kurzweil’s Move to Google Will Accelerate More Human-like (and Humanistic) Virtual Agents
Ray Kurzweil is a gifted prognosticator when it comes to the timing of technological breakthroughs. In July 2011, in this interview with a Wall Street Journal correspondent Andy Jordan, Kurzweil pegged 2029 as the year that computers will routinely pass the Turing Test – referring to “a test of a machine’s ability to exhibit intelligent behavior, equivalent to or indistinguishable from, that of an actual human.” Natural language processing and understanding has long been the yardstick for measuring the human-like qualities of computers, so it is no surprise that Kurzweil, in his capacity as Director of Engineering, will reportedly focus on machine-learning and language-processing projects.
We are entering a year when top design shops like Frog Design have declared that the “Human-computer interaction gets more humanistic,,” or as Creative Director Mark Rolston put it “We need to take the computers out of computing and more humanistically integrate them into everyday life.” By hiring Kurzweil, Google plans to be at the center of such developments. And given the competitive nature of the markets that Google is serving – search, mobile apps, e-commerce – a lot will happen well in advance of the vaunted 2029 (assuming we all survive the Mayan Apocolypse on December 21, 2012).
Google’s interest in Kurzweil’s work is long standing. Larry Page, Google’s co-founder, helped Kurzweil and Dr. Peter Diamandis set up Singularity University in 2008. Over the years, Google has supported it with more than $250,000 in donations, and some of Google’s earliest employees are among the university’s “founding circle,” meaning that they have made personal donations of $100,000 each. Courses at Singularity U address challenges and accelerate development around nanotechnology, artificial intelligence, energy, biotech, robotics and computing.
Conflating “Human-like” with “Humanistic”
“Human-like” speaks for itself. We’re already witnessing some very impressive platforms for human-to-machine conversations. They include sufficiently accurate speech recognition and elegantly sculpted text-to-speech rendering, coupled with computer resources that use natural language processing and elements of “artificial intelligence” to make sense out of what someone has said by keeping track of context and managing a polite dialogue. These systems are also “adaptive,” in that they learn from mistakes and are constantly improving. The current level of quality for “virtual assistants” (so far a succession of female personas like Siri, Lola, Lexee and a few others) have given a broadening number of companies confidence to incorporate them into their customer care workflows.
A search on Google shows that “humanistic” means “of or pertaining to a philosophy asserting human dignity and man’s capacity for fulfillment through reason and scientific method.” That’s quite different from “human-like” but it is certainly a worthy objective and one, I would argue, that might capture the role that Kurzweil is destined to play at Google. Inasmuch as the whole Singularity movement is defining the many ways that a combination of technology and humanity results in a sum that is greater than its parts, we’re talking about both man and machine’s capacity for fulfillment through reason and scientific method.
The race to build the best Personal Virtual Assistant should shorten the time it takes for a system to pass the Turing Test – at least at the language understanding level. Meanwhile we should all benefit as Google, Apple, SRI, Microsoft and a number of smaller firms keep improving on the concept. At the same time, companies like Interactions Corporation are applying their technology to make self-service both more human-like and more humanistic. The result is a software platform that has evolved into a “virtual assistant” that is very accurate in recognizing what callers are saying and quickly routing them to the correct resource. It is human-like, in that it understands what a callers are saying with great accuracy and its responses sound like live agents. It is humanistic because it targets the simple objective of helping callers complete their tasks quickly and efficiently.
2013 will witness much higher quality person-to-machine interactions as developers pursue a parallel course that continue to make them more human-like and humanistic.