Explainable AI: Do people understand machine decisions better when they can interact freely with AI-based systems?