White, Black, Brown and Metallic: Do Robots Have a Race and Why It Matters
If you search Google Images for "robot," you will find mostly white or metallic devices. The same is true in movies, whether you watch "Blade Runner," "Star Wars" or "The Terminator." Nearly all robots in films look white or metallic.
One could assume that a white robot blends in better with its surroundings. Or that it simply looks nicer, since it does not absorb as much light as a darker robot. Or that we picture robots in a bright future, and white robots look bright. According to a recent study, the reality may be much darker. People prefer white robots because racism exists in the world. Let us look at examples of experiments involving robots of different colors.
Robots and Race
Some will think this is going too far and argue that robots are not people. Others will say that no sensible person would treat a robot like a human, because people are not that foolish. Unfortunately, this is far from the case. Anthropomorphism, the practice of creating objects that resemble people, is common throughout robotics. This principle also assumes the human mind's ability to see human traits in everything it interacts with.
Previous research has already shown that people tend to project gender stereotypes onto machines. If you ask Siri or Alexa what gender they are, they will say they have no gender. Yet their voices sound female to the human ear. This is done deliberately: according to several studies, people prefer to hear male voices when it comes to authority, but want to hear a female voice when they need help.
Evidence is now emerging that people prefer white robots over other colors. Of course, this preference does not concern the robots themselves. At least at this stage of history, robots have no consciousness.
The real problem is what biases this reinforces in people. However inconvenient robots may be, their capabilities in many areas exceed those of humans. They perform repetitive tasks without complaint, work with incredible precision and need no vacations or days off. Now imagine that all or most of these robots are white.
In our subconscious thinking, we may associate the traits of white robots with light-skinned people. This association is not only unjustified but may also reinforce white supremacy in the broader culture. Studying how people see robots helps mitigate such scenarios.
Black Robots and the Use of Force
Christoph Bartneck and his colleagues at the University of Canterbury in New Zealand are among the pioneers of human-robot interaction research. In a productive study from 2018, they found that the shooter bias paradigm applies not only to dark-skinned and light-skinned people but also extends to robots.
The paradigm works as follows: study participants were instructed to shoot an agent as quickly as they could if it had a weapon. However, they should not shoot if the agent was holding a harmless object, such as a phone, a wallet or a can of soda. The experiment was conducted digitally. Participants saw a photo of the agent and pressed a button to shoot.
Previous studies showed that, regardless of the participants' race and ethnicity, they tended to shoot dark-skinned agents faster and more often than white ones. Bartneck and his colleagues showed that the same bias holds for black and white robots. For this experiment, they used the Nao robot and photoshopped its color to match the typical skin tone of a dark-skinned or a white person. They then mixed images of these black and white robots holding weapons or other objects with images from the original study. They presented all of this to study participants and found that they reacted to black robots the same way they reacted to dark-skinned people.
This implies that participants attribute some race or ethnicity to black robots. When the participants were asked what race the robots were, even though a "no race" option was available, more than 90% chose a race, and about 70% of participants believed the color of the robots corresponded to a race.
In the future, when robots of all races walk through cities, people may cause more harm and deaths to black robots. At least this could happen if people remain as they are today.
Brown Robots Level the Playing Field
In a follow-up study in 2019, Bartneck and his colleagues wanted to test how reliable their results were. In particular, they wanted to find out whether the results were related only to social priming, whether the bias extended to other races or ethnic groups, and whether robots that looked more or less human could change the results in any significant way.
Some readers of the original study pointed out that the findings could be related to social priming. This is a phenomenon in which people, when given information of a certain type, allow that information to influence their subsequent actions. During the experiment, researchers asked participants about their race and nationality, and this could have subconsciously affected their later behavior. However, when they repeated the experiment without asking participants these questions beforehand, they still found the same racial bias in the results.
Another critical point was that the original study included only black and white robots, which is a gross oversimplification and erases dozens of other races. Therefore, they included a brown robot in later versions of the study. Surprisingly, with these three robots in the test, the shooter bias disappeared. Participants began shooting at all robots at the same speed and accuracy, regardless of their appearance.
The study's authors interpreted this result as unexpected evidence of the value of diversity. However, this is still just one study, and so far the only one of its kind. The researchers had an acceptable number of participants and agent images: each of 160 people was shown 60 images.
The researchers included other robots in the experiment. The Inmoov robot looks somewhat more human than the original Nao robot, while Robosapien looks more like a machine. However, participants' performance did not change depending on how human the robot was. When asked about this, participants said they did not perceive one robot as more human than another.
Making robots more human-like does not change people's attitudes toward these devices. However, their color leads people to perceive some type of race and show the same biases they show toward other people. Whether the presence of more than two different races for robots eliminates these biases remains to be confirmed in future research.
A Less White Future
If robots can have a race, can they be biracial or multiracial? This would at least reduce the likelihood that we end up with the stereotypical, very white robots we know from many films. It would also help us avoid another dystopian scenario, in which robots of all colors exist, but the authoritative and powerful ones are white, while the obedient and servile ones are colored.
If multiracial robots become widespread, both of these dystopian scenarios become less likely. This idea is somewhat inspired by the recently presented home robot Amazon Astro, which checks for leaks and intruders when the owner is away, serves drinks and entertains guests. Astro is a robot with a screen and black front parts, as well as white wheels and body.
The recently announced Tesla Bot from Elon Musk follows the same direction: its head, neck and shoulders are black, while the rest of its body is white. To some, its appearance resembles a light-skinned woman in a black headdress. Thus, it may also be a statement against religious prejudice.
These two recently presented robot models are not going to magically erase all the terrible human biases we have. It is also unclear whether other robot manufacturers will follow this example and focus their designs on less white models. But they offer an interesting alternative and inspire some hope for a future with less bias and bigotry.
Only time will tell how and when we eliminate bias from robots, artificial intelligence and other technologies. But it seems that major technology companies are aware of the problem.
