This chapter investigates the critical dimension of Artificial Intelligence Literacy, focusing on the evaluation of AI systems and their outputs. It first outlines the primary challenges to critical understanding, beginning with the need for advanced data and information literacy to navigate the risks inherent in generative AI—such as bias, misinformation, and hallucinations. It then examines the black-box problem, establishing explainable AI (XAI) and user-centred techniques like Chain-of-Thought (CoT) prompting as fundamental conditions for trust and accountability. To address these challenges, this chapter proposes two corresponding pedagogical strategies. To foster critical information use, it details an Inquiry-Based Learning (IBL) approach that turns tools like ChatGPT into objects of critical investigation. To demystify AI’s internal logic and decision-making, it presents a Problem-Based Learning (PBL) framework, employing practical activities such as error analysis and the construction of simple decision trees. This chapter concludes that these hands-on, inquiry-driven methods are essential for moving learners from passive consumption to active, critical, and responsible engagement with AI technologies.