It seems very hard for OpenAI to prove that independence. Since they seem unable to exclude the possibility that their model was trained on transcripts by the two mathematicians.