Is It Safe to Put Personal Information into ChatGPT? | What Not to Enter, and How to Mask It Before Sending

Drafting an email to a customer, summarizing inquiry history, tidying up a contact list. The jobs you most want to give ChatGPT are the ones full of names and contact details. Many of you have probably stopped and wondered, "Is it really okay to paste this in as is?"

This article is for people who use ChatGPT at work, and for people in IT and general affairs who set the company's rules for using generative AI. We go through, in order, how the information you enter is handled inside ChatGPT, what Japan's Personal Information Protection Commission has warned about, what you should not enter, and how to mask it when you really need to use it.

On how ChatGPT handles data, we only describe what OpenAI's official site and Help Center said as of October 26, 2026. Terms can change, so check the latest official pages before you use it. How Gemini handles data is covered in Is It Safe to Put Personal Information into Gemini?.

Bottom line: don't enter personal information, and if you must, mask it first

  • On individual plans (Free, Plus, Pro), what you enter may be used for training under the default settings. You can stop training under Data Controls in the settings
  • On business plans (Business, Enterprise, and so on), what you enter is not used for training by default. If your company uses ChatGPT, a business plan is the baseline
  • Even on plans that don't train on your data, what you enter is sent to OpenAI's servers and processed there. "Not used for training" does not mean "not sent"
  • In June 2023, the Personal Information Protection Commission warned that if you enter prompts containing personal information, you should thoroughly confirm that it is within the scope of the purpose of use and that the provider does not use it for machine learning
  • In practice, the most reliable approach is to mask names, contact details, and ID numbers before sending. Even with them masked, AI can handle many tasks, such as summarizing, organizing tables, and drafting messages

How ChatGPT handles the information you enter

Individual plans (Free, Plus, Pro)

According to OpenAI's Help Center, on the individual services (Free, Plus, Pro), what you enter may be used to train models. To stop this, turn off "Improve the model for everyone" under Data Controls in the settings. Once it's off, conversations from then on are not used for training. Your saved conversation history stays. The setting applies to the whole account.

For conversations you don't want kept in your history, you can also use Temporary Chat. Temporary chats are not saved to your history and are not used for training. However, OpenAI notes that a copy may be kept for up to 30 days for safety purposes.

Business plans (Business, Enterprise, and so on)

For ChatGPT Business, Enterprise, Edu, and the API, OpenAI officially states that it does not use organizational data to train models by default. Some plans let the organization set the retention period.

However, this is a promise about "not training." What you enter is processed on OpenAI's servers, so you are still handing personal data to an outside service. You need to decide separately how much is acceptable to enter within your company.

The Personal Information Protection Commission's alert (June 2023)

On June 2, 2023, the Personal Information Protection Commission issued an alert on the use of generative AI services. Its points for businesses were these two:

  • When entering prompts that contain personal information, thoroughly confirm whether doing so is within the scope necessary to achieve the specified purpose of use
  • If you enter prompts containing personal data without the person's consent and that data is handled for purposes other than producing the response, it may violate the Act on the Protection of Personal Information. If you do enter it, thoroughly confirm, among other things, that the provider does not use that data for machine learning

For general users, it listed the risk that entered personal information is used for machine learning and output as inaccurate content, the possibility that responses contain inaccurate personal information, and the need to decide after checking the terms of use and privacy policy. On the same day, it also issued an alert to OpenAI, the provider of ChatGPT, about the collection of sensitive personal information and the notification of purposes of use.

In other words, you are expected to check both "Did I set it so my data isn't used for training?" and "Do I need to enter this at all?"

Information you should not put into ChatGPT

When drawing the line in company rules, it's easier to decide if you split information like this.

Category Examples How to treat it
Information that identifies a person Name, address, phone number, email address, date of birth As a rule, don't enter it. If needed, mask it first
Numbers and financial information My Number (Japan's individual number), credit card numbers, bank account numbers Don't enter it
Sensitive personal information Medical history, beliefs, criminal record, and so on Don't enter it
Credentials Passwords, API keys, access tokens Don't enter it. If it leaks, replace the key itself
Confidential information Unpublished figures, contract details, transactions with customers Follow company rules. A business plan is the baseline

It's tempting to think "one person's worth is fine," but with a contact list or inquiry history hundreds of rows long, even the person pasting it hasn't looked at everything in it. Oversights usually happen when a large batch is pasted at once.

Real leak incidents and their causes are collected in Generative AI Data Leak Cases.

What to decide as a company

  • Decide which services and plans may be used: use business plans for work, and don't handle work information in personal accounts
  • Decide where the line is on what may be entered: set it by category as in the table above, and share it with examples
  • Check the training settings: if individual plans remain in use, have everyone check their Data Controls settings
  • Put a check in place before anything is entered: rules alone can't prevent an oversight at the moment of pasting

How to mask personal information before sending

Replace it by hand

Replace names with "Person A" and email addresses with "(email)" before pasting. You don't need any tool, but with many entries it takes time, and some replacements get missed.

Phrase the request so it works while masked

For summarizing, organizing tables, or writing email text, you rarely need the personal information itself. If you ask in a masked form, such as "a thank-you email to [Name]," the AI writes just the text. You fill in the final details on your end.

Use a browser extension that checks the moment you paste

There are also browser extensions that check what you paste or attach on generative AI sites and mask it before sending. The advantage is that you don't rely on checking by eye, and large numbers of rows can be checked at once. When adopting one, check who provides it and whether it sends the content outside for inspection.

Checking before sending, inside the browser: Samurai App Data Guard

Samurai App's Chrome extension, "Samurai App Data Guard," checks the content before it is sent when you paste text or attach a file in ChatGPT, Gemini, or Claude.

  • What it finds: names, email addresses, phone numbers, addresses, dates of birth, My Number IDs, credit card numbers, bank account numbers, API keys and tokens, and passwords. It finds names using clues such as table headers and labels like "Name:"
  • What you can choose: cancel, send as is, send partly masked, or send masked. When masked, each match is replaced with its category name, such as [NAME] or [EMAIL]
  • The check stays in the browser: it uses no AI model and checks with fixed rules. It never sends the content outside for inspection

Even with information masked, ChatGPT can still organize tables and draft messages. The flow from pasting a customer list into ChatGPT to sending it masked is shown in the video on the PII detection page (45 seconds). For details, see the PII detection page.

Sources

Related featurePII detection and masking before data reaches AIPII detection finds email addresses, phone numbers, My Number IDs, API keys and more in the files and text you send to AI, right in the browser, and lets you send a masked copy. No model is used and nothing is sent out for the check.

More in this category

Back to AI and security articles

Everyday work comes together, connects, and becomes the company's strength.

Contact usOpen Samurai App

See the demoContact us