4. Research, Articles and Insights
Our articles and insights discuss research in AI safety, interpretability, verification and related fields, and report on developments reported by others. We work to make them accurate and to cite our sources, but research moves quickly: findings are revised, methods are superseded, and a piece reflects what was known when it was published.
Examples, figures and product illustrations on the Site that are labelled as illustrative are exactly that. They show how something works and are not reported results from any client or system.