SAFE, PRIVATE AND FAIR

BY THE END YOU CAN

• Know what never goes into a chat and what happens to your chats
• Spot hidden instructions and approve actions safely
• Know what to do after a mistake, and keep decisions about people with people

7 short lessons · about 7 minutes · 4-question quiz.

IN THIS CHAPTER

1. What not to paste
2. Your chats and your data
3. Follow your company's rules
4. Hidden instructions in content
5. Approve before it acts
6. If something goes wrong
7. Fair to people

Plus a 4-question quiz with every answer explained.

GO DEEPER: FURTHER READING

Hand-picked guides, articles and videos on this chapter's topics. Links checked 2026-11-06.

• GUIDE
  Anthropic Privacy Center

  Anthropic's answers on how your data is used, kept and deleted, for personal and company plans.

• REFERENCE
  Usage Policy (Anthropic)

  What Anthropic does not allow Claude to be used for. Legal wording.

• REFERENCE
  Prompt injection (OWASP)

  Security experts explain hidden instructions, with real examples. Some technical words.

• GUIDE
  Use Cowork safely (Claude Help Center)

  Safety advice for letting Claude act in your files and apps.

• REFERENCE
  AI Risk Management Framework (NIST)

  For managers and policy owners: the US government's guide to managing AI risk in an organisation.

SOURCES · CHECKED 2026-11-06

• Updates to consumer terms (Anthropic)
• Anthropic Privacy Center
• Use connectors (Claude Help Center)
• Use Cowork safely (Claude Help Center)