What is Generative AI?

For most of machine learning's history, models did discriminative tasks: classify this email as spam or not, predict whether this loan will default, detect objects in this image. The model observed an input and produced a label or prediction. Generative AI inverts this: given a prompt or partial input, the model produces completely new content that didn't exist before. The breakthrough came from models that learned the underlying distribution of their training data well enough to sample new examples from it. Language models like GPT generate text by predicting each next token in sequence. Diffusion models like Stable Diffusion and DALL-E generate images by iteratively denoising random noise into coherent pictures. Music models like Suno generate songs from text descriptions. Video models like Sora and Veo generate clips from prompts. Code models like Claude Code write and edit entire codebases. What makes generative AI transformative is that creating content — the fundamental activity of knowledge work — is suddenly automatable in ways it never was before. First drafts of essays, marketing copy, product descriptions, code scaffolding, design mockups, meeting summaries can all be produced in seconds. The human role shifts from creating from scratch to editing, directing, and curating AI-generated drafts. This is why generative AI is reshaping knowledge work faster than any previous technology wave.