Responsible AI must minimise violations on users’ legal rights