Speaker presents dangers of artificial intelligence, calls for algorithm accountability
Cathy O'Neil, author of "Weapons of Math Destruction," spoke on how unchecked algorithms may produce and reinforce classist, racist and sexist results in hiring practices, policing and sentencing. O'Neil encouraged students to be wary of blind faith in the internet.
Cathy O’Neil, a pioneer of algorithmic auditing and author of the 2016 book “Weapons of Math Destruction,” presented at Pellegrene Auditorium on Tuesday night for the 2023 Norman L. Ford Science Literacy Lecture Series. The series brings in renowned scientists each year to discuss the newest advancements on a range of scientific topics and the impact of science in our daily lives. O’Neil began the discussion with a statement that was reinforced time and again in the presentation:
“Algorithms don’t just predict the future, they cause it.” O’Neil said that AI is intimidating and difficult to understand—something she pointed out is by design.
She defined AI in simple terms, explaining that algorithms are based on past history of success. The hazard of algorithms is defining what “success” means.
“Success is determined by the people in power,” O’Neil said.
This creates a skewed power dynamic between users and controllers of algorithms. Further, those most harmed by algorithms are also individuals who are disenfranchised in other parts of their lives. O’Neil used the acronym WMD to describe the subset of algorithms that are dangerous and unfair: widespread, mysterious and destructive.