🎧 In this episode of The Ginsbourg's Podcast, Season 2, Episode 1, hosts Shay Ginsbourg and AI co-host Omer delve into the critical and rapidly evolving field of testing AI-based software systems. Drawing insights from Shay Ginsbourg's paper, "Testing AI-Based Software Systems: From Theory to Practice," this discussion navigates the fundamental paradigm shift from traditional, deterministic software testing to the complex, probabilistic, and adaptive nature of AI systems. The episode meticulously explores five core testing dimensions unique to AI: Data Integrity and Quality, Non-Deterministic and Adaptive Behavior, Bias and Fairness Testing, Explainability and Transparency, and Robustness, Security, and Monitoring. Listeners will gain a comprehensive understanding of the challenges posed by AI's black-box nature and data dependency, alongside practical methodologies such as Model-Based Testing, Data-Driven Testing, Adversarial Testing, and Explainable AI. The discussion also addresses the landscape of standardization and regulation, including the ISO/IEC TR 29119-11:2020 and FDA guidance for AI in medical devices, highlighting both their significance and limitations. This episode is an essential listen for graduate-level audiences and professionals seeking to ensure the reliability, trustworthiness, and ethical deployment of intelligent systems in an AI-driven world.