Cravath Publishes “Tech Explainer” on Diffusion Models

In March 2025, Cravath continued its “Tech Explainers” series, which introduces and explains technical, cutting‑edge concepts and describes their importance in a way that is accessible to and easily understood by business and legal professionals outside the technical community.

The new installment in the series discusses, at a technical level, how generative AI models are trained to produce images from text prompts through “diffusion” and how image generation works at the inference stage. The article highlights how, as generative AI cases move through the courts and the U.S. Copyright Office refines its position on copyright eligibility, it is imperative that legal practitioners understand the underlying technology in order to make informed assertions and adequately advise clients and stakeholders.

Please click here to read the full article.