A new court filing alleges that Meta "took precautions" to not seed copyrighted data after downloading 82 TB dataset ...