There were a number of A.I.-related accidents in 2018 and 2019, but only a few made the headlines. An Uber self-driving car hit and killed a pedestrian in Tempe, Arizona—but there were countless more incidents that didn’t result in death, and as a result, aren’t known to the public.

At the moment, researchers are not obligated to report accidents or incidents involving our data, or A.I. processes, unless a law is broken. While big companies must inform consumers if their personal data—credit card numbers, home addresses, passwords—have been stolen, they are not required to publicly document instances in which algorithms have learned to discriminate against someone on the basis of race or gender, for example.