AI Comedy of Errors, Or: The Importance of Choosing Your AI Performance Metric With Care

Using the correct measure of performance is one of the first steps in assessing the possible impact of an AI system. But which performance metric is "the correct one"? Is it accuracy? The answer, as always, depends on the context in which the AI system will be used. In this blog post, we take a closer look at what accuracy means in machine learning, and, using the example of a "hate speech detection" AI system, examine which other performance metrics might sometimes be more appropriate.