Tip:
Highlight text to annotate it
X
"I think I saw a person"
"It seemed very real"
"Come on, it could be a man"
Hello, my name is Eva
What is your name?
Eva is a robot built by Hanson Robotics
It has 32 degrees of freedom to represent 48 muscle actions of the human face
It also has a skin made from structured porosity elastomer material
which allows simulating human-like facial expressions
It can generate various movements and speak
But can it be considered an intelligent robot?
It depends.
People have a different perceptions of robots
"quite scary" ... "pretty, pretty robot"
and different expectations
"I hear the word robot and...
imagine a machine with a lot of cables"
"It was not what I had in mind"
But above all a lot of people attribute different human qualities to robots
"It was just kind of shaking its head when I was wrong"
"It was nodding when I was right"
"The robot moves the eyes according to the words"
"It always turned to me when I was speaking and..."
"closed the eyes a little bit as if it was focusing on me"
The robot doesn't shake its head to say yes or no
The eye movements are not synchronized with the generated speech
and it doesn't follow the users' movements with its eyes
What the robot does is generate idle head and eye movements
synchronize lip movements with the speech
and simulate some facial expressions
The interaction between the robot and users is bounded by a script
The entire conversation and robot behavior are pre-programmed
Both the robot and users should follow the script while having a conversation
All possible questions and answers are displayed on the computer screen
or on the tablet PC
"Can you tell me what is the capital city of Switzerland?"
"It's Bern" "Correct answer, bravo"
"Bern" "Correct answer, bravo"
But what happens if we try to go beyond the script?
"I'd love it to be Geneva, but"
"Sorry, wrong answer. Was it a difficult one for you?"
One of the most common attributions people make to robots is intelligence
Some people attribute mental skills to the machines
that go much beyond the script and software
It doesn't matter whether the robot is inherently intelligent
but rather whether the robot appears intelligent
"I almost forgot it was a robot"
"The robot was able to understand me"
"It looks like the robot is not...eh"
"only pre-programmed"
"but sometimes when the person asked"
"eh, made another comment"
"the robot knew how to answer to that"
One can adopt the strong artificial intelligence stance
That is, assume that it is possible to duplicate human intelligence
in artificial systems
On the other hand one can follow the stance of weak artificial intelligence, as we do
where human intelligence is only simulated rather than duplicated
This is where the human tendency to project intelligence onto robots becomes useful
It is always humans who set benchmarks for the artificial intelligence
and decide whether the robot is intelligent or not
The way people perceive robot intelligence depends on different factors
Above all robot design and the individual tendency to anthropomorphise the machines
Besides, things that matter are the physical and social environment in which robots are
being used and the task they are supposed to accomplish.
What seems to be an intelligent action in one case
might seem as not so intelligent in another case
"The robot knew how to change the subject"
"I think its response every time is exactly the same"
In any case, the illusion of intelligence in a robot can be achieved through
projective intelligence on the part of the human observer.
The difficulty lies in maintaining the illusion over time