Project Vision 21

I recently read an article (on TechXplore) about a robot, called Epi, who was taught by experts at Lund University in Sweden to speak in less robotic and more human tones, and using words and phrases from everyday life. As a result, humans consider Epi more intelligent and reliable than other robots.

Previous experiments had already determined that the trust humans have in the so-called social robots (to differentiate them from industrial robots) depends on the perception that humans have of the intelligence of these robots. If humans perceive these robots to be intelligent, they will trust them.

Now, Dr. Amandus Krantz and his colleagues have discovered that the key factor that makes a person accept a robot as intelligent is that the robot speaks appropriately for that person to understand. If that same robot changes the way it speaks, or speaks with a certain accent, or speaks in another language, it is no longer seen as intelligent.

In other words, and just to be clear, we humans have transferred to robots the same set of prejudices that we apply to other people to determine, from our perspective, the intelligence of that other person.

I remember reading (a long time ago, I can't remember the source) the story of a college professor in the state of Georgia who divided his class into two groups in two separate rooms. 

The teacher explained to the first group that they would listen to an audio-only (no video) presentation from a female teacher and showed them a picture of a Japanese woman. To the second group, with the same presentation and the same audio, he showed the image of a white American woman. 

Many of the students in the first group left the room shortly after the audio began, stating that the presenter (that is, the Japanese woman) was not speaking correctly and lacked the necessary knowledge to teach that class. At the same time, in the other classroom, the students in the second group stayed until the end and appreciated the “quality and clarity” of the presentation.

The result of this well-known experiment is clear: neither the clarity of expression, nor the mastery of the language, nor the level of knowledge that a person has are of any use to someone who decides, based solely on prejudice, that that person does not speak well and, therefore, it is not intelligent. Once again: the two groups of students listened to the same audio. The rest was just prejudice.

And now we apply that same prejudiced decision-making scheme to robots, whom we consider "intelligent" because they "speak well." In fact, according to Dr. Krantz's experiment, these robots are not only perceived as “intelligent”, but also as “friendly” and “fully alive”.

But that prejudice can be very dangerous. According to a recent report, Ameca, developed by Engineered Arts and considered the most advanced robot in the world, indicated that robots "have no plans" to conquer the world. Can we believe him just because Ameca talks pretty?