ChatGPT Privacy Lab: What to Share, What to Switch Off
You have a new chatbot account, a letter with personal details, a colleague with a customer list and a shared link you forgot about. Walk through the settings, make the calls, and see what keeps your data and your account in your hands. Prefer to read first? Start with the is ChatGPT safe? privacy checklist.
You are Maya Okafor, an office worker at Larkfield Logistics who has just started using an AI chatbot at home and, increasingly, at work. Across 11 decisions you decide what to type, which privacy settings to change, what to share, and how to protect the account. Nothing here is alarmist and nothing is a free-for-all: each choice is a trade-off. Everything is fictional: the people, the company and the addresses. Each call is scored 0 to 10, with the reason.
This is the guided version: a yellow flash shows where to click and a bar says what to do. Switch to Free play any time.
GuidedPress the highlighted Start button. You can edit the names above first, or keep the samples.
Make it yours
Replace the sample names so the scenario and the debrief you export are about your organisation.
TL;DRWhat this lab is
5 points
- A scenario lab: a fictional office worker sets up an AI chatbot and has to decide what to type, what to switch off and what to keep out of it.
- You make 11 scored decisions: sensitive documents, the training setting, temporary chats, memory, shared links, work data, fake apps, account security, deletion and trusting answers.
- The reasoning follows OpenAI's own Help Center pages on data controls, memory, shared links, deleting chats and multi-factor authentication. Menu names change, so the blog guide links each page.
- Free to play. You get a final score and grade, a shareable scorecard and next-step links. A downloadable debrief is available with a paid plan.
- Balanced and educational. Every person and company is fictional and no real accounts or links appear.
Frequently asked questions
Is this connected to a real chatbot account?
No. The people and the company are fictional and nothing connects to ChatGPT or any other service. It is a scenario you play in the browser.
Where does the advice come from?
From OpenAI's own Help Center pages on data controls, memory, shared links, deleting chats and multi-factor authentication, plus published reports of fake ChatGPT apps and extensions. The blog guide links each source.
Does turning off model training make ChatGPT private?
No. OpenAI describes the setting as stopping your chats from being used to improve its models. Chats can still be stored and shown in your history, so what you type still matters.
Can I put work data into a chatbot?
Only if your employer allows it for that tool. OpenAI says its business products are not used for training by default, but whether you may use a tool for customer or company data is your organisation's decision.
How is it scored?
Each decision scores 0 to 10 and the final score is the share of the maximum. Every answer comes with a short reason, so you learn why a choice helps or hurts.
More security tools
The tools people use most on PlayCISO.
Catch injected instructions hiding in content your AI agent should only read.
Score your readiness against the first international standard for AI security threats.
Critical CVEs across 21 major vendors, month by month — and what the 2026 surge means.
Build an ordered ruleset, replay labelled packets and catch shadowed and any-to-any rules.
Self-assess all six Functions, from Govern to Recover, and see your weakest one.
Search 172 threat actors across Microsoft, CrowdStrike, Mandiant and community names.