Large language models (LLMs) and applications like ChatGPT capture our imagination with their human-like interactions. But their prowess hinges on a fundamental truth: these models are only as insightful, unbiased and effective as the data that trains them, and the data on which they are applied. We will look at the different ways in which data, good and bad, can influence and be influenced by LLMs, considerations such as responsible and regulated AI, and a strategy for ensuring quality data provides the foundation for the voice of generative AI.