The point is that no human had access to the same data, because the LLM used the unpublished work of a number of mathematicians. The only way for a human to gain access to the same work, i.e. the same dots to connect, is to be in active collaboration with those mathematicians. In which case they would all be included in the resulting paper.