Is ChatGPT Learning From Your Clients' Data? Check This Setting

Your clients sign a privacy agreement. Here is the ChatGPT setting that decides whether your chats train the model, what business plans do by default, and why a chat can end up in court.

Your clients sign a privacy agreement. It says you won't share their private information, and that includes the AI tools you type into every day.

I covered this in my AI class at the South Bay Association of REALTORS on October 28. It's the setting I'd check first, and most agents I know have never opened it.

## The setting

On a personal ChatGPT plan (Free, Plus or Pro), OpenAI can use your conversations to train its models unless you turn that off. On a computer, open Settings, choose Data controls, and switch off "Improve the model for everyone." On your phone, tap the profile icon, then Settings, then Data controls.

Turning it off doesn't delete or hide your saved chats. It stops your chats from being used for training.

Voice mode has its own training control, separate from the text setting. If you talk to ChatGPT out loud about a deal, check both.

## Business plans are different

In class I put it this way: "In the business ones, they're turned off automatically." OpenAI's policy for ChatGPT business and enterprise plans says it doesn't use your inputs and outputs to train its models by default.

So find out which one you're on. A personal login on a work laptop is still a personal plan.

## Claude works on an opt-in basis

Anthropic's privacy center says consumer Claude chats are used to improve Claude only when you choose to allow it. Incognito chats are not used for that, even if the setting is on. Look at the setting in your own account either way.

## Chats can end up in court

Courts are the other half of this. In the New York Times copyright case against OpenAI, a federal court ordered the company to keep ChatGPT output logs, including some that users had deleted. Later, a federal judge ordered a sample of 20 million de-identified logs produced. The broad preservation order was narrowed that fall. Check the court docket for where the case stands today before you rely on these details.

A chat sitting on a company's servers is a record, and records can be ordered produced in a lawsuit. I'm a broker, not an attorney, so this is not legal advice. Ask your broker or your own attorney what your files have to keep, and for how long.

## My rule for the chat box

Assume anything you type could be read back to you someday. Leave out names, addresses, income, account numbers and anything else that identifies a client. If you wouldn't forward it to a stranger, don't paste it.

Every tool from the class, with its link and the free notes, is on one page: huntermason.com/ai20.

Tony Self

Harcourts Hunter Mason Realty, CA DRE 01906720

Sources (checked October 28, 2026. These pages change, so check them before you rely on them):

• OpenAI Help Center, Data controls in ChatGPT: https://help.openai.com/en/articles/7730893-data-controls-in-chatgpt

• OpenAI Help Center, ChatGPT Enterprise and Team data policy: https://help.openai.com/en/articles/8983130-what-is-the-chatgpt-enterprise-and-team-data-policy

• OpenAI Help Center, Voice mode FAQ: https://help.openai.com/articles/8400625

• Anthropic Privacy Center, Is my data used for model training?: https://privacy.claude.com/en/articles/10023580-is-my-data-used-for-model-training

• Court orders in New York Times v. OpenAI (MDL No. 25-md-3143), as reported by https://terms.law/2025/11/12/openai-v-new-york-times-stopped-being-just-a-copyright-case-the-moment-the-court-turned-to-your-chatgpt-logs/ and https://engadget.com/ai/openai-no-longer-has-to-preserve-all-of-its-chatgpt-data-with-some-exceptions-192422093.html

ChatGPT, Claude and OpenAI are trademarks of their respective owners. This post is not affiliated with or endorsed by any of them.