# ChatGPT: Keep Client Data Safe, Set Up Your Team's Data Controls in 10 Minutes

> A client asks if their files go into ChatGPT. Set data controls, memory, Temporary Chat and shared links, then copy a one-page team AI policy.

Source: https://learn.ynteractive.com/content/chatgpt-data-controls-protect-client-data · Updated 2026-10-13 · Free from Stepthrough (https://learn.ynteractive.com)

## Practice it step by step

[Start the free practice](https://learn.ynteractive.com/modules/chatgpt-data-controls)

You do every step yourself in a practice copy of ChatGPT. It checks your work as you go, and “Do it for me” finishes any step you get stuck on. Free, about 10 minutes, and nothing touches your real accounts. [Open full screen](https://learn.ynteractive.com/modules/chatgpt-data-controls)

Free · Live now · [AI Basics](https://learn.ynteractive.com/content/course/ai-basics) · [ChatGPT](https://learn.ynteractive.com/content/tool/chatgpt) · ~10 min · 16 steps · Beginner

Updated October 13, 2026. Built against the tool’s current screens and checked step by step. [How we check this lesson](https://learn.ynteractive.com/content/about#how-we-check)

A client emails your firm: do our files ever go into [ChatGPT](https://learn.ynteractive.com/content/tool/chatgpt), and what settings do you use? If nobody can answer in a sentence, you have a client confidentiality problem, not an AI problem. In this free, hands-on lesson you play the office manager of a six-person bookkeeping firm, set the ChatGPT controls that matter in about ten minutes, and finish with a one-page team AI policy you can copy.

## At a glance

| Question | Answer |
| --- | --- |
| What you build | A client asks if their files go into ChatGPT. Answer with confidence: set the data controls, memory, Temporary Chat and shared links, then hand your team a one-page AI policy. |
| Tools | ChatGPT |
| Time | About 10 minutes, 16 steps in 3 parts |
| Level | Beginner, no coding needed |
| Cost | Free |
| Ways to learn it | [Practice it step by step](https://learn.ynteractive.com/modules/chatgpt-data-controls) |

### The 3 parts, in short

1. The business case: Read the client’s question, Read Priya’s note to herself.
2. Set the controls: Open ChatGPT, Open your account menu, Open Settings, Turn off Improve the model for everyone, and 8 more steps.
3. Write the team policy: Open the team policy, Copy the policy.

**Start here first:** [AI Agent vs Chatbot vs Automation: One Job Done Three Ways](https://learn.ynteractive.com/content/ai-agent-vs-chatbot-vs-automation) (module 1 of AI Basics)

**In the course:** [Part 3 · ChatGPT for Everyday Work](https://learn.ynteractive.com/content/course/part-3-chatgpt-for-everyday-work) (Module 13 of the [full course](https://learn.ynteractive.com/content/course))

**Part of:** [AI at Work: Start Here](https://learn.ynteractive.com/content/paths/ai-at-work-start-here) (step 2 of 10). Next up: [Prompt Writing 101: Role, Context, Task, Format](https://learn.ynteractive.com/content/prompt-writing-101-role-context-task-format)

## The business case: one paste can cost you a client

Hartwell Bookkeeping is a made-up six-person firm. Its client, Delgado Dental, asks in writing whether payroll and patient billing files are ever put into ChatGPT, because their insurer now asks every vendor. At the same time, the office manager knows that last week a new bookkeeper pasted a client payroll export into ChatGPT on a personal login, and nobody knows which settings that account has.

That is the pattern in most small firms: no malice, no policy, and personal accounts with default settings. A vague answer can cost a renewal. A clear one takes about ten minutes of settings and one page of rules. This lesson covers the settings. For what should never be pasted in the first place, start with [Safe AI at Work](lesson:safe-ai-at-work).

## Step 1: turn off Improve the model for everyone

Click your name at the bottom left, choose Settings, and [Data controls](https://learn.ynteractive.com/content/glossary/chatgpt-data-controls) is the page you want. The first real decision is the switch next to [Improve the model for everyone](https://learn.ynteractive.com/content/glossary/improve-the-model-for-everyone), described on screen as "Control whether your chats are used for model training". On personal plans, switching it off stops new chats from being used to train [OpenAI](https://learn.ynteractive.com/content/glossary/openai)'s models. Do this on every account anyone uses for work. OpenAI says business plans such as ChatGPT Business and Enterprise do not train on your business data by default, so confirm the terms for your plan. See also [ChatGPT data controls](term:chatgpt-data-controls) in the glossary.

Turning it off does not delete your history and does not make chats secret. They are still processed to give you an answer, so the habit of removing names and numbers still matters.

## Step 2: review Shared links, Export data and Delete all chats

On the same page, Shared links has a Manage button. Its dialog says anyone with a shared link can view the shared chat, and that deleting a link stops others from viewing it but does not delete your original chat. A link to a client conversation can sit there for months, so put a monthly review in the calendar.

Lower down are Archived chats, Archive all chats, Delete all chats and Export data. Export data is how you keep a record, and Delete all chats removes everything. A sensible firm rule: export before you delete, and only the account owner deletes. Your page may show extra rows depending on your plan.

## Step 3: decide on memory

Under Settings, Personalization, Enable memory lets ChatGPT personalize answers from your chats, files and connected apps, and Memory summary has a Manage button to see what it has kept. On a personal login that is useful. On a shared work login it can carry one client's details into another client's chat, so most firms turn it off for shared accounts and review the memory summary on any account that stays on.

## Step 4: use Temporary Chat for sensitive work

At the top right is the Temporary Chat icon. The notice says a temporary chat ignores memory, plugins and custom instructions and won't appear in your history. The line at the bottom adds that, for safety, OpenAI may keep a copy for up to 30 days. So it is safer, but not secret: it does not replace removing names, account numbers and other personal details first. See [Temporary Chat](term:temporary-chat).

## Step 5: write the one-page team AI policy

Settings only work if everyone follows the same ones. The lesson ends with a copyable one-page policy: one approved account (a business plan if you can); training off; memory off on shared logins; Temporary Chat for sensitive work; never paste passwords, full bank or card numbers, Social Security numbers, health details or confidential client files; review shared links monthly; export before you delete; and ask the policy owner when unsure. Adapt the names and your own contract terms. This is not legal advice, and your client contracts and local rules come first.

## What this saves you

About ten minutes of setup replaces an open-ended scramble every time a client, insurer or auditor asks how you use AI. You get one approved account, a written answer you can send the same day, and a team that knows the rules. Next business case: [reply to client emails faster in ChatGPT](lesson:chatgpt-reply-faster), with these settings already in place.

## How the interactive lesson works

Each step is shown in the Watch panel with a short explanation, then you do it yourself in Your turn on a practice copy of ChatGPT. No account is needed, no setting is changed and nothing is sent. The firm, the client and the email are made up.

## Links

- [OpenAI Help Center](https://help.openai.com): OpenAI's own help pages for data controls, Temporary Chat and memory. Plans and options change, so check your own screen.

## Try it yourself, free

Reading the steps is a start. Doing them is how it sticks. The interactive module walks you through every click in a practice copy of ChatGPT, checks each step, and never touches your real accounts.

[Start “ChatGPT: Keep Client Data Safe, Set Up Your Team's Data Controls in 10 Minutes”](https://learn.ynteractive.com/modules/chatgpt-data-controls)

## Step by step: ChatGPT: Keep Client Data Safe, Set Up Your Team's Data Controls in 10 Minutes

All 16 steps of the interactive module, in order. In the module you watch each one, then do it yourself in a practice copy that checks your work.

### 1. The business case

1. **Read the client’s question.** Read the email from Delgado Dental. Click the **question** about their files.
   Priya runs the office at Hartwell Bookkeeping, a six-person firm. A client, Delgado Dental, has just asked in writing: are our payroll and patient billing files ever put into ChatGPT, and what settings do you use? Their insurer is asking them to check.
2. **Read Priya’s note to herself.** Click the **note** below the email.
   And here's why Priya's stomach drops. Last Tuesday a new bookkeeper pasted a client payroll export into ChatGPT, on her personal login. Nobody had told her not to, and nobody knows which settings that account has. A wrong answer to the client could cost you the renewal, or worse. A right answer takes about ten minutes.

### 2. Set the controls

3. **Open ChatGPT.** Click the **ChatGPT** tab.
   Let's set it up. In real life, do this on the account your firm approves, signed in as the person who manages it.
4. **Open your account menu.** At the bottom left, click your name, **Priya**.
5. **Open Settings.** Click **Settings**.
6. **Turn off Improve the model for everyone.** You’re on **Data controls**. Click the switch next to **Improve the model for everyone** to turn it off.
   Settings opens on Data controls. The first control that matters is Improve the model for everyone. On personal plans it decides whether your chats can be used to train OpenAI's models. Do this on every account anyone uses for work. Business plans don't train on your data by default, which is one reason firms pay for them.
7. **Look at your Shared links.** Click **Manage** next to **Shared links**.
   Next, Shared links. When someone shares a chat, anyone with the link can read it. That's handy for a team, but a link to a client conversation can sit there for months.
8. **Close the list.** Read the note, then click the **×**.
   Notice what the window says. Deleting a shared link stops others viewing it, but it doesn't delete your original chat. Put a monthly check in the calendar: open this list, and delete any link nobody needs. Close it.
9. **Find Export data.** Click **Export** next to **Export data**. In this practice copy it only shows a hint.
   Two buttons further down are for record keeping. Export data emails you a file of your chat history. Delete all chats removes everything. The firm's rule: export first if you might need a record, and never use delete all without a reason. In the practice copy, Export only shows a hint.
10. **Open Personalization.** In the left menu, click **Personalization**.
    Now memory, which lives under Personalization.
11. **Turn off Enable memory.** Click the switch next to **Enable memory** to turn it off.
    Enable memory lets ChatGPT personalize answers from your past chats. On a personal login that's useful. On a shared work login it's a risk: one client's details could resurface in another client's chat. So for shared logins, the firm turns it off. Underneath, Memory summary has a Manage button, to see or clear what ChatGPT has kept.
12. **Back to the chat.** At the top left, click **Back to app**.
    That's the account settings done. Go back to the chat.
13. **Try Temporary Chat.** At the top right, click the **Temporary Chat** icon.
    One more tool, for the moment you do have to work on something sensitive. At the top right is Temporary Chat.
14. **Read what it promises.** Read the notice. Click it when you’re done.
    Read the notice carefully. A temporary chat ignores memory, plugins and custom instructions, and it won't appear in your history. But look at the line at the bottom: for safety, OpenAI may keep a copy for up to thirty days. So it's safer, not secret. It still doesn't replace taking out names and numbers first.

### 3. Write the team policy

15. **Open the team policy.** Click the **Team AI policy** tab.
    Settings only work if everyone follows the same ones. So write them down.
16. **Copy the policy.** Read the eight rules, then click **Copy policy**.
    Eight short rules: one approved account, training off, memory off on shared logins, Temporary Chat for sensitive work, never paste the dangerous things, review shared links monthly, export before you delete, and ask when unsure. Then reply to Dr. Delgado with it. For what not to paste, see the Safe AI at Work lesson.

## Frequently asked questions

**How do I stop ChatGPT from training on my chats?**

Go to Settings, then Data controls, and switch off Improve the model for everyone. It applies per account, so check every account your staff use for work.

**Is Temporary Chat private?**

Safer, not secret. It ignores memory and stays out of your history, but ChatGPT's own notice says a copy may be kept for up to 30 days for safety.

**Should I turn ChatGPT memory off at work?**

On shared work logins, most firms do, so one client's details never surface in another client's chat. On a personal login, review the Memory summary and decide.

**What are ChatGPT shared links and how do I remove them?**

A shared link lets anyone with it read a chat. Settings, Data controls, Shared links, Manage lists them, and deleting a link stops others viewing it without deleting your chat.

**What should a small firm's AI policy say?**

At minimum: which account to use, which settings must be on or off, what must never be pasted, who owns the policy and who to ask. The lesson ends with a one-page version to copy.

**Do ChatGPT business plans train on my data?**

OpenAI says it does not train on business data from plans such as ChatGPT Business and Enterprise by default. Confirm the terms for your exact plan and follow your contracts.

[Start this module free](https://learn.ynteractive.com/modules/chatgpt-data-controls)

- Previous in AI Basics: [Same Task, Four Assistants: ChatGPT vs Claude vs Gemini vs Copilot](https://learn.ynteractive.com/content/chatgpt-vs-claude-vs-gemini-vs-copilot-same-task)

## More in AI Basics

- [AI Agent vs Chatbot vs Automation: One Job Done Three Ways](https://learn.ynteractive.com/content/ai-agent-vs-chatbot-vs-automation): Chase six unpaid invoices three ways: a chatbot writes a template, an automation follows a rule, and an AI agent works out the steps and flags the tricky one.
- [What Is a Webhook? Send a Test and Watch the Data Arrive](https://learn.ynteractive.com/content/what-is-a-webhook): Point a plumber's website contact form at a free webhook tester, send a test and a real message, and watch the data arrive the moment it's sent.
- [Safe AI at Work: What Not to Paste, Training Settings and Checking Answers](https://learn.ynteractive.com/content/safe-ai-at-work-what-not-to-paste): Turn off model training, strip a tenant's bank details and address out of a prompt, then catch the two promises ChatGPT made up before the reply goes out.

[All 5 modules in AI Basics](https://learn.ynteractive.com/content/course/ai-basics)

## More ChatGPT tutorials

- [Prompt Writing 101: Role, Context, Task, Format](https://learn.ynteractive.com/content/prompt-writing-101-role-context-task-format): Ask ChatGPT for the same customer email twice, once vaguely and once with a four-part prompt, and see exactly why structure gets a better answer.
- [Get Your Business Recommended by ChatGPT](https://learn.ynteractive.com/content/get-recommended-by-chatgpt): Ask ChatGPT who it recommends in your town, see which pages it trusts, then use it to write an FAQ page and Business Profile copy that answer real customer questions.
- [Summarize a Meeting Transcript with ChatGPT](https://learn.ynteractive.com/content/chatgpt-summarize-meeting-transcript): Paste a long meeting transcript into ChatGPT and get the decisions, an action items table with owners and due dates, and a follow-up email ready to send.
- [Reply Faster: Turn a Messy Customer Email Thread into a Clear, Friendly Reply](https://learn.ynteractive.com/content/chatgpt-reply-to-customer-emails): Turn a messy customer email thread into a clear, friendly reply: give ChatGPT the thread and your facts, check every promise, fix the tone, and send it from Gmail.
- [ChatGPT Projects: Keep Your Brand Voice, Files and Instructions in One Place](https://learn.ynteractive.com/content/chatgpt-projects-brand-voice): Set up a ChatGPT Project with your brand voice instructions, your price list and your style guide, so every new chat already knows your business.
- [Analyze a Spreadsheet with ChatGPT: Upload a CSV, Get a Chart and Three Takeaways](https://learn.ynteractive.com/content/chatgpt-analyze-spreadsheet-csv): Upload a CSV of sales to ChatGPT, have it check the data first, ask your questions, get a chart you can download, and finish with the three takeaways that matter.

## From the blog

- [Team AI Policy: What Staff Can Paste Into ChatGPT](https://learn.ynteractive.com/content/blog/team-ai-policy-what-staff-can-paste-into-chatgpt): A one-page team AI policy for a small firm: which account to use, which settings to switch off, and what staff must never paste into ChatGPT.

[All blog posts](https://learn.ynteractive.com/content/blog)

## Terms in this lesson

Plain-English definitions, each with an example from a free lesson.

[ChatGPT data controls](https://learn.ynteractive.com/content/glossary/chatgpt-data-controls) · [Temporary Chat](https://learn.ynteractive.com/content/glossary/temporary-chat) · [Improve the model for everyone](https://learn.ynteractive.com/content/glossary/improve-the-model-for-everyone) · [OpenAI](https://learn.ynteractive.com/content/glossary/openai) · [Custom instructions (ChatGPT)](https://learn.ynteractive.com/content/glossary/custom-instructions)

[The full AI glossary](https://learn.ynteractive.com/content/glossary)

## Browse related tutorials

[AI Basics (5)](https://learn.ynteractive.com/content/course/ai-basics) · [ChatGPT tutorials (33)](https://learn.ynteractive.com/content/tool/chatgpt)