This podcast discusses whether the AI mechanisms reinforce and amplify things that have been wrong in the past, with certain machine learning approaches potentially having serious problems due to the data that has been given.