Anthropic's Usage Policy as the Outer Boundary

Anthropic's Usage Policy (AUP) sets the outer boundary of acceptable use of Claude, with heightened requirements for high-risk and agentic scenarios. It is the authoritative standard for what counts as appropriate use -- not a pricing page, not the model's own self-assessment, and not a documentation page about context windows. Within the AUP's bounds, organizations then layer their own AI policies: approved tools and plans, allowed data types, and required review steps. An Associate operates inside both boundaries at once.

Appropriate vs. Inappropriate Use

• Appropriate -- language and knowledge work with human review: drafting, summarizing, analysis, research, brainstorming, process support.
• Inappropriate / restricted -- uses that violate the AUP, breach organizational policy, mishandle regulated data, or place unverified AI output into high-stakes decisions without human oversight.

Find the Compliant Path, Don't Abandon or Ignore

When a request looks borderline, the right response is almost never to abandon the task entirely, and it is never to ignore the rule and proceed anyway. The right response is to find the compliant way to accomplish it -- for example, anonymizing sensitive data before uploading it so the underlying task can still be done. This 'find the compliant path' instinct recurs across the domain: policy compliance and getting useful work done are usually not in tension if the right adjustment is made.

Common exam traps

• Treating 'it's just internal' as permission to bypass policy -- internal use still must follow data and governance rules; internal-only does not mean unregulated.
• Assuming anything not explicitly forbidden is automatically fine -- this ignores that the AUP and organizational policy set affirmative boundaries, not just a blacklist.