Visual analytics for explainable and trustworthy machine learning

AI systems can solve complex problems, such as navigating a vehicle through heavy traffic or diagnosing a patient's illness upon admission to a hospital.

However, people need to trust that machines will perform their intended tasks, which is difficult with AI because it lacks simple and deterministic execution guarantees. As a result, establishing trust in AI systems is crucial.

To increase trust, visual analytics can be utilized, where interactive graphs and charts enable users to observe how the AI makes decisions and even to influence the process. Therefore, it is imperative for practitioners, machine learning experts, and users to monitor and regulate AI's decision-making processes, understand how it works, and prevent it from producing unexpected results.

The challenge is to develop strategies that will foster trust between humans and machines.