Generative AI

What Generative AI is

Generative models learn the distribution of their training data well enough to sample plausible new examples from it. The same underlying idea covers a chat assistant writing an email and a diffusion model painting an image.

How it works

Most generative systems are trained with self-supervision on very large corpora, then adapted with instruction tuning and preference tuning. At inference the model samples from a probability distribution, which is why the same prompt can yield different results.

Why it matters

Generative AI moved AI from behind-the-scenes prediction into everyday creative and knowledge work, and it is the category most organisations are budgeting for today.

Common uses

• →Drafting, editing and summarising written work
• →Image, video and audio production
• →Code generation and refactoring
• →Synthetic data and scenario generation