//
Search across all documentation pages
6 pages in this section.
Introduces the mental model behind Claude's safety guardrails, from training choices to usage limits.
A starting guide to Claude's built-in safeguards, content limits, and everyday responsible-use habits.
Explains why Claude's training data has a fixed end date and how that affects answers.
Explains why Claude.ai answers from training data alone unless a browsing feature is enabled.
A checklist for verifying outputs, protecting sensitive data, and setting team-wide usage guardrails.
Enumerates practices for using Claude safely, verifying outputs, and respecting Anthropic's usage policies.