I'm sure detailed mouse and keystroke data can actually leak health data from subjects. What are the odds one can detect early parkinson disease from mouse wiggle data? If such data leaks away health status, I think capture should be forbidden under current rules.