Computer vision research datasets have been criticized for violating subjects’ privacy, reinforcing cultural biases, and enabling questionable applications. But regulating their use is hard.
Yet the data that has been taken down for ethical reasons isn’t just available — it continues to be used prominently in academic research. Further, even if a dataset was created for benign purposes, it may have uses in more questionable areas. Oftentimes, these uses are enabled by a derived dataset.