AI literacy is the baseline knowledge someone needs to work sensibly with artificial intelligence: how AI tools broadly produce their outputs, where they tend to fail, and how to judge when to rely on them. In practice, that means knowing a large language model (LLM) predicts plausible text rather than looking up facts, that its answers can vary between runs, and that its output reflects the data it was trained on.
So what? Testers and quality people meet AI on both sides of their work: as tools that help them test, and as features inside the products they test. Without a baseline understanding of how those systems behave, it is hard to decide what an AI tool should be trusted with, or what to look for when testing one. In the EU, the AI Act has required organisations that provide or deploy AI systems to ensure sufficient AI literacy among their staff since February 2025.
Examples: Treating a single passing run of an AI feature as provisional, because the same input can produce a different answer next time. Checking company policy before pasting customer data or proprietary code into a public AI chatbot. Explaining to a product owner why an AI coding assistant suggested a deprecated library function: its training data predates the change.
