A VIRTUAL, AI-rained robot has displayed shocking signs of racism and sexism, fueling rising fears of a technology takeover, 报道说.
A study released last month shows that new robots trained with artificial intelligence have exhibited biases that could prove to be extremely harmful.
Institutions including John Hopkins University and the Georgia Institute of Technology released a 学习 last month arguing that “robots enact malignant stereotypes.”
这 研究 shows that artificial intelligence algorithms tend to show biases that could unfairly target people of color and women in their programs of operations.
In a recent experiment, scientists tasked virtual robots with sorting billions of pictures with related captions.
The robots repeatedly paired the word “criminal” with pictures of a Black man’s face.
The robots also reportedly associated words like “homemaker” 和 “janitor” with pictures of women and people of color.
Researcher Andrew Hundt said: “The robot has learned toxic stereotypes through these flawed neural network models.”
添加: “We’re at risk of creating a generation of racist and sexist robots.
“But people and organizations have decided it’s OK to create these products without addressing the issues.”
Most read in news tech
The researchers found that their robot was 8% more likely to pick males for each task. It was also more likely to pick white and Asian men.
Black women were picked the least out of every category.
The primary concern is that biased robots like this could enter homes.
然而, researchers hope that companies will work to diagnose and improve the technological issues that led to the harmful biases.
Researcher William Agnew of University of Washington added: “While many marginalized groups are not included in our study, the assumption should be that any such robotics system will be unsafe for marginalized groups until proven otherwise.”