AI Bias Can Be Fixed. And There Is a Tooth Fairy
June 22, 2018
AI algorithms were supposed to remove the bias and human error from many decisions, from loan applications to criminal sentencing. As with most techno-fixes, experts soon discovered that algorithms have unforeseen flaws, like how some AI can accidentally create more bias. We learned one way this is being corrected from a recent Fast Company story, “This Tool Lets You See—And Correct—The Bias in an Algorithm.”
According to the article:
“The tool uses statistical methods to identify when groups of people are treated unfairly by an algorithm–defining unfairness as predictive parity, meaning that the algorithm is equally likely to be correct or incorrect for each group. ‘In the past, we have found models that are highly accurate overall, but when you look at how that error breaks down over subgroups, you’ll see a huge difference…’”
This is an interesting tool. We need to be able to correct these biases since most experts agree that we are also the cause of many of them. Finding a solution to this ever-growing problem provides hope not only for future use of AI technology, but also for financial investment in these technologies.
Has anyone seen the tooth fairy today?
Patrick Roland, June 22, 2018
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
