AI Testing Glossary: Key Terms Every QA Engineer Should Know
AI is changing how software is built, and more importantly, how it’s tested. If you’re a QA engineer working with machine learning models, LLMs, or intelligent systems, traditional testing vocabulary only gets you halfway there. The real challenge? Understanding the new language of AI testing. This glossary breaks down the most important terms you’ll encounter in modern QA environments, explained in plain English, with context you can actually use in your day-to-day work. Why AI Testing Needs Its Own Vocabulary Unlike traditional applications, AI systems are non-deterministic . That means the same input doesn’t always produce the same output. So instead of verifying fixed outputs, QA engineers now evaluate probabilities, patterns, and behaviors . That shift introduces new concepts, many borrowed from data science, some unique to testing AI-driven systems. Core AI Testing Terms Every QA Should Know 1. Model Accuracy What it means: The percentage of correct predictions made by a ma...