Global Themes In AI Data: Notes From The Road

To protect against the potentially catastrophic results of AI bias, we must train models on the right inputs.