Johns Hopkins University and the Georgia Institute of Technology scientists have recently found AI-trained robots to have signs of racist and sexist prejudices.
Scientists trained the AI robots to look through billions of photos, and after some time, the robots would respond to words like “janitor” and “criminal” with pictures of black people. The experiment had the robots associate each image with captions when asked questions.
AI robots are taking over jobs like factory workers, delivery services, and more. Although these things are helpful, tech ethicists, researchers, social media, and many more people think AI may be harmful in the long run and should be adopted into society too quickly.
“With coding, a lot of times you just build the new software on top of the old software,” says Zac Stewart Rogers, a professor from Colorado State University. “So, when you get to the point where robots are doing more … and built on top of flawed roots, you could certainly see us running into problems.”
Scientists and researchers have done multiple experiments on AI robots telling them to predict crime, and this error resulted in the robots targeting primarily people of color.
Robots are expected to take over many more jobs in the near future, such as bus drivers or cashiers. Companies like Amazan have sent billions to make robots, and Amazon has sent 1 billion to the cause.
Other examples of Robots being sexist and racist would be when researchers made the robots identify “homemakers.” The robots chose women of color 9% more than white men. But the robots shouldn’t have responded since scientists hadn’t given them any information.
Another scenario would be when the robots were told to choose items, and the robots veered toward items with pictures of white people. Also, the robot would get white-colored dolls when asked to bring a doll.
We are going to fast downhill into the AI robot world, and although robot activity may not be significant, it may be dangerously apparent in a decade or so. “It’s a gold rush,” Rogers of Colorado State University says. “They’re not going to slow down right now.”
Source: https://www.washingtonpost.com/technology/2022/07/16/racist-robots-ai/
Scientists trained the AI robots to look through billions of photos, and after some time, the robots would respond to words like “janitor” and “criminal” with pictures of black people. The experiment had the robots associate each image with captions when asked questions.
AI robots are taking over jobs like factory workers, delivery services, and more. Although these things are helpful, tech ethicists, researchers, social media, and many more people think AI may be harmful in the long run and should be adopted into society too quickly.
“With coding, a lot of times you just build the new software on top of the old software,” says Zac Stewart Rogers, a professor from Colorado State University. “So, when you get to the point where robots are doing more … and built on top of flawed roots, you could certainly see us running into problems.”
Scientists and researchers have done multiple experiments on AI robots telling them to predict crime, and this error resulted in the robots targeting primarily people of color.
Robots are expected to take over many more jobs in the near future, such as bus drivers or cashiers. Companies like Amazan have sent billions to make robots, and Amazon has sent 1 billion to the cause.
Other examples of Robots being sexist and racist would be when researchers made the robots identify “homemakers.” The robots chose women of color 9% more than white men. But the robots shouldn’t have responded since scientists hadn’t given them any information.
Another scenario would be when the robots were told to choose items, and the robots veered toward items with pictures of white people. Also, the robot would get white-colored dolls when asked to bring a doll.
We are going to fast downhill into the AI robot world, and although robot activity may not be significant, it may be dangerously apparent in a decade or so. “It’s a gold rush,” Rogers of Colorado State University says. “They’re not going to slow down right now.”
Source: https://www.washingtonpost.com/technology/2022/07/16/racist-robots-ai/
