How does AI learn?

What does it mean when someone says an AI has learnt something?

Sandra Wachter is Professor of Technology and Regulation at the Oxford Internet Institute, where she researches the legal and ethical implications of AI and regulation of online platforms.

[doctorf]

As the lady says that the way to perpetuate prejudices

[carrellwashington]

Perhaps, in simply trying to evaluate subjective and objective patterns in LLM to the point of profiling, you've introduced bias because you're profiling... Aside from the promoting positive evaluations, what factors truly contribute to the furtherance of just, moral, law practices? Or, what is objectively counterproductive to the practice of law? Maybe the avoidance of these characteristics is another method of elimination. As an aside, all of these assume law like behavior (read causality and prediction). What if the rejected candidate codifies a political and social movement in legal terms that is, in all respects, more advanced than the ones currently in place? What does AI have to say, then?