As the EU AI Act becomes broadly applicable in 2026, trust in AI stops being optional - it becomes a legal requirement. High-risk systems in healthcare, finance, human resources (HR), automotive, and the public sector will need more than just working code. They will need proof: evidence of safety, transparency, and fairness. And that evidence will come from testers.
This will change the game for QA leaders and testers. Testing is not only about functionality or performance anymore - it is about being audit-ready, compliant, and trustworthy.
In this hands-on tutorial, Usha Kandala will strip away the jargon of the EU AI Act and show attendees how to turn its requirements into practical testing strategies. Participants will dive into real case studies - classifying systems by risk, designing test approaches for bias and robustness, and preparing the kind of documentation regulators expect. The session will also explore tools like Fairlearn, SHAP, and audit checklists that make compliance testable and repeatable.
Attendees will leave with their own AI Compliance Test Readiness Kit, including a risk classification worksheet, a bias-testing demo, a sample conformity checklist, and clear strategies for building test evidence that regulators will actually accept.
Whether attendees are already testing AI systems or just getting ready for what is coming, this session ensures that “testing at its best” means testing that not only finds bugs, but also builds trust.