Users don’t understand computer explanations for image labeling errors

When images are uploaded to online platforms, they are often tagged with automatically generated labels that indicate what is shown, such as a dog, tree or car. While these labeling systems are often accurate, sometimes the computer makes a mistake. Providing explanations to users to interpret these mistakes can be helpful, or sometimes even necessary. However, researchers at Penn State’s College of Information Sciences and Technology found that explaining why a computer makes certain mistakes is surprisingly difficult.