I think stuff like this is only acceptable, morally and perhaps legally, if they tell you what they uncover / discover about you with this. That way we can ver

I think stuff like this is only acceptable, morally and perhaps legally, if they tell you what they uncover / discover about you with this. That way we can verify it's all above board. Imagine they put the "eye movements / mouth movements / head movements" through a machine learning system and discover that certain head movements lead to poorer results (or maybe not poorer results, but slightly slower results). Only it turns out these particular head movements, etc. are due to a disability protected by the americans with disabilities act. They may not have fed this into the system directly, but if the decision was made based solely (or primarily) on these things, are they not breaking the law? And how could it even be traced in this case if the person doesn't yet know they are going to have said disability, but the ML algorithm somehow learned that the person is likely to (for example) become deaf within the next few years.