Ray Kurzweil’s AI Net is a phenomenal newsletter and site that report on scientific breakthroughs, speculation about the future, and the thoughts of eccentric genius Ray Kurzweil.
Today’s newsletter reported on Virtual Eve: first in human computer interaction
PhysOrg.com, Nov. 19, 2007
A near-human virtual teacher called Eve can tell if a child is frustrated, angry or confused by an on-screen teaching session and can adapt the tutoring session appropriately.
With a human-sounding voice, Eve can ask questions, give feedback, discuss questions and solutions and show emotion. To develop the software, the Massey University researchers observed children and their interactions with teachers and captured them on thousands of images.
For decades, a test proposed by Alan Turing in his 1950 publication Computing Machinery and Intelligence has been the gold standard for measuring computer “intelligence.” The test: given two chat sessions, one written by a human and the other by a computer, can an inquisitor tell which is which?
As the world’s worldview transitions from absolutism (yes or no?) to pragmatism (shades of gray), I propose a new test: Is it good enough that people will respond to it as if it were human? If Eve can read a student’s emotions and respond in kind, that’s enough. I don’t have to believe Eve is real to respond to her as if she were human.
Related posts:
{ 2 comments… read them below or add one }
I have mixed feeling about this. When I place a call and am required to interact with a machine, I expect the efficiency of a human and when I don’t get it, feel frustrated, a human response. When I know the correct responses – that is, how to use the phone touchpad to get the quickest response from the ‘intellegent system’ I applaud the programmers and my own user skill and ability to get what I need without the inefficiency of dealing with a human.
As a technical services call center rep, I pride myself on the use of human response and quick analysis to solve most problems more efficiently that any ‘smart systems’ available to our customers.
In the future, smart systems will increasingly be part of the medical diagnostics field. It remains to be seen how long this will take to be welcomed by the public, and fine tuned to make it as effective (and friendly?) as dealing with a human person.
John, as I understand Eve, the point is that I needn’t be a touchpad expert for the system to help me. I welcome smart systems that understand what I need.
Anyone else reading Don Norman’s latest book, which focuses on the relationship between people and machines?