Computers that learn words from texts written by humans capture their meaning – and also figure out biases, according to a new study published in the journal Science.
"Machines can learn word associations from written texts and . . . these associations mirror those learned by humans," the researchers wrote, noting, for example, artificial intelligence uncovered the association between "pleasantness and flowers or unpleasantness and insects."
"It can also tease out attitudes and beliefs," the researchers found. "Such biases may not be expressed explicitly, yet they can prove influential in behavior."
The researchers found, for example, European-American names were more closely associated with words like "honest," "gentle," though unpleasant words – like "divorce," "jail" – were more likely to be attributed to African-American names.
Young people were considered pleasant, old people were not, they found.
And AI associated women names more with family and the arts than with mathematics.
"Artificial intelligence learns biases, but it needs the awareness not to make prejudiced decisions," Princeton researcher Aylin Caliskan said in a podcast about the research. "Since machines do possess self-awareness the way humans do, a human in the loop can help machines make ethical decisions."
© 2025 Newsmax. All rights reserved.