Last reviewed
Correct answer: B. Use an Incognito chat, which is excluded regardless of that setting
Explanation
The principle — data handling has both an account-level control and a per-conversation one, and knowing the second saves you from toggling the first.
Why the key is correct — Anthropic states that Incognito chats are not used to improve Claude even if Model Improvement is enabled in your Privacy Settings. That is precisely the case described: a one-off conversation you want excluded, without disturbing how everything else is handled.
Why the others are wrong — deleting afterwards is not the documented mechanism for exclusion. A request typed into the conversation is just more conversation and changes nothing about how it is processed. And turning the account setting off would work, but treating it as the only option means changing a standing preference for a single chat.
Remember this — Incognito is the per-conversation control; the privacy setting is the standing one.
Sources — Anthropic's article on whether your data is used for model training.
Sources
“Your Incognito chats are not used to improve Claude, even if you have enabled Model Improvement in your Privacy Settings.”
Practise 10 questions on this topic
Take Claude.ai Basics — Timed Test 1 (10 questions) — scored instantly, explanation for every question, no login.