Do explanations for data-based predictions actually increase users' trust in AI?