Assignment 4.4 Building Trust in AI: Explainability, Validation, and Performance Metrics

The infographic explains the concept of Explainable AI (XAI) and why it is important for building trust in modern AI systems such as GPT, Claude, Gemini, and LLaMA. It shows that explainability helps people understand how AI models make decisions, which improves transparency, accountability, and fairness. The graphic also highlights key challenges, such as the complexity of large models, the lack of clear reasoning in outputs, and the difficulty of meeting legal and ethical requirements. In addition, it presents validation methods like cross-validation, red teaming, and human evaluation, along with performance metrics such as accuracy, precision, recall, perplexity, and hallucination rate. These methods and metrics help measure how reliable and safe AI systems are.

The design of the infographic follows a clear structure that moves from explanation to evaluation and then to improvement. This flow helps show how explainability, validation, and performance metrics are connected and how they work together to improve AI systems. Color-coded sections and simple visuals were used to make the information easy to understand and well organized. The inclusion of current techniques, such as attribution methods, interpretability tools, and governance practices, shows that improving explainability is an ongoing process. Overall, the infographic highlights that strong validation and clear performance measures are necessary to ensure that AI systems are trustworthy, effective, and suitable for real-world use.