Short answers, pulled from the story.
Firmin Didot coined the word stereotype in 1798 to describe a metal printing plate that could duplicate any typography. This mechanical invention was designed to save time and money in the printing trade by allowing publishers to mass-produce identical copies of a page without resetting the type.
American journalist Walter Lippmann gave the word stereotype its modern psychological meaning on the 2nd of May 1922 in his book Public Opinion. Lippmann argued that the human mind creates simplified pictures of reality to make sense of the chaos, serving as cognitive shortcuts that allow people to navigate social interactions with speed and efficiency.
The stereotype content model developed by Susan Fiske and colleagues organizes stereotypes along two primary dimensions: warmth and competence. Warmth refers to how friendly or trustworthy a group appears, while competence measures how capable or successful they seem, creating four distinct emotional responses based on the intersection of these dimensions.
Psychologist Patricia Devine demonstrated in 1989 that racial stereotypes are activated subliminally, meaning that even individuals who consciously reject prejudice still experience the automatic activation of negative associations when exposed to cues related to a stereotyped group. Her experiments showed that words related to the cultural stereotype of Black people were presented subliminally, and subjects rated a race-unspecified target person as significantly more hostile if they had been primed with racial words.
Claude M. Steele and Joshua Aronson conducted the first experiments showing that stereotype threat can depress intellectual performance on standardized tests, with black college students performing worse than white students when the task was framed as a measure of intelligence. The threat of confirming a stereotype creates a mental burden that undermines performance, leading to a cycle of anxiety and underachievement that is difficult to break.
A 2024 study testing how AI text generators respond when asked about African Americans revealed that while the program would respond with positivity when asked directly, it would generate results reminiscent of the Jim Crow era when asked in a more roundabout way about people who speak African American English. This suggests that AI systems are not neutral arbiters of truth but are instead mirrors reflecting the biases present in their training data.