How to Opt Out of ChatGPT Using Your Training Data While Keeping Chat History
ChatGPT defaults to using your chat history and chat interactions as training data for the ChatGPT service and AI model. One of the most obvious reasons for this is that prior interactions with ChatGPT can be used to refine the Large Language Model and to improve the service, but there are obviously some privacy and personal ramifications here too.
Essentially this means your chats and interactions with ChatGPT are not really private, since it’s possible that machine learning, AI, or perhaps even people, could be viewing and using your chat history and engagement with the chatbot LLM to train and improve the model.
If you have ever dug around in the ChatGPT settings for privacy and data controls, you’ve likely have noticed there’s a toggle for turning off the training data usage, however it comes with also disabling your ChatGPT chat history. This leads many users to think that you can’t have ChatGPT chat history while also opting out of using your chats as training data. But it turns out that’s not entirely true.
With a little-known privacy toggle that’s buried within the OpenAI privacy page, you can opt out of using your chats as training data, while simultaneously keeping your ChatGPT chat history enabled. We’ll show you how to do this.
How to Stop ChatGPT Using Your Content as Training Data But Keep Chat History Enabled
We’re going to show you how to make a request to OpenAI to disable the use of your content as training data, while still preserving the handy Chat History feature. Here’s how to do this:
- Go to ChatGPT as usual and login to your OpenAI account
- Now in a web browser, go to this page at privacy.openai.com/policies
- Click on “Do not train on my content – ask us to stop training on your content”
- Enter the email address you use to login to ChatGPT to submit the request
- Check your email, and again login to your ChatGPT account to complete the request
It may take some time for the request to be completed, but ChatGPT and OpenAi will (likely) approve your request and your content and chats will no longer be used for training data.
You must perform this request and action when using ChatGPT directly from the OpenAI website or the ChatGPT app on iPhone or iPad.
You can not complete this request if you’re accessing ChatGPT through Microsoft Edge (as many do to access GPT 4 for free) or through Bing (which many also use for GPT 4 access), or through another service that uses ChatGPT, unless it’s associated directly with our email anyway.
Whether or not you want to make this change is up to you, and how you use ChatGPT, and how personal or private your chats and conversations are. If you’re spilling the beans about very personal things to ChatGPT, you might want to disable the usage of your content as training data, but if you’re just geeking around with the LLM and exploring with nothing particularly private or personal, keeping the chat content as training data feature may be fine too. It’s really up to you and your individual preferences.
It’s important to remember that almost nothing is truly private online in this day and age, with any online service. That realization may be enough to turn some into luddites, while others just tread a bit more cautiously with their data online, and others throw all caution to the wind and air it all out there for every service and global alphabet agency to sniff. It’s all a matter of individual preferences, but just be aware that the days of genuine true privacy online are largely gone, whether you think you’re being sneaky or not – VPN, TOR, proxy, or otherwise.
