Text this: Establishing and evaluating trustworthy AI: overview and research challenges