Have you ever wondered how AI models justify their decisions? Large language models (LLMs) can now generate human-readable explanations, but a new study reveals they don't always explain things clearly. Researchers explored whether LLMs can adapt their explanations to different levels of understanding, from sixth grade to college level. They found that while LLMs can adjust their writing style, traditional readability metrics don't always accurately capture the complexity. Interestingly, explanations with "medium" complexity—like those aimed at high schoolers—often correlated with higher quality ratings, perhaps because they strike a balance between detail and clarity. However, human readers in the study didn't always perceive these differences as intended, suggesting our understanding of how humans and AI process information might need further exploration. This research opens up exciting new avenues for understanding how to make AI more transparent and accessible. Could fine-tuning LLMs on different types of explanations dramatically improve how AI interacts with users in the future? This is just the beginning of unraveling the complexities of explainable AI.