Disparate impact applies to actions which are unjustified. For instance imagine I decide to never hire somebody who likes rap music. That would have a disparate

Disparate impact applies to actions which are unjustified. For instance imagine I decide to never hire somebody who likes rap music. That would have a disparate impact on a certain group of people, yet since it probably has nothing to do with the job I'm hiring for, it would be unjustified and could be argued to be a form of disparate impact discrimination. By contrast imagine for a labor job I decide to never hire anybody who can't left and carry at least 120 pounds. That would also have a disparate effect negatively impacting a protected class, but it would be justified. Machine learning takes all data and draws conclusions that map strongly against the data and ideally generalize to new scenarios. So long as the behavior trying to be predicted for was relevant to the task at hand, using the recommendations of such algorithms would certainly be justified and thus not fall under disparate impact. In a nutshell disparate impact is not to ensure equality of result, but to prevent discrimination by proxy. And discrimination not being of the form 'I'm not going to hire women because they can't lift as much as men' but of the form 'I'm not going to hire women because I don't like working with women.'