Why it matters
On personal accounts, ChatGPT and Claude can use your conversations to train future models. Both have a setting that stops it, and switching it off in both apps takes about five minutes.
Get the rest of this guide free
Enter your email once and every guide and setup on the site opens, this one included. You'll also get AI This Week by Southbeam: the week's AI news for business owners, every Friday.
Steps in ChatGPT
- On the web, open your account menu and select Settings.
- Select Data controls.
- Select Improve the model for everyone, turn it off, and select Done.
- On iPhone or Android, open the sidebar, tap your profile icon to open Settings, tap Data controls, and turn off Improve the model for everyone.
- While you're there, check the audio setting. OpenAI's dictation help page says you can stop sharing voice recordings by turning off Include your audio recordings, or by turning off Improve the model for everyone entirely.
When you're signed in, the setting follows your account across devices. From then on, your new conversations won't be used for training. Your saved chats stay where they are.
Steps in Claude
- On the web or desktop, select your name, then Settings.
- Select Privacy.
- Turn off the model training toggle. Anthropic's help page calls it Help improve our AI models.
- On mobile, the path is the same: your name, then Settings, then Privacy, then the toggle.
Anthropic says that once it's off, it won't use your new chats and coding sessions for future training, and it stops using your older stored chats in future training runs. Training runs that had already started still include them. This setting covers Claude Free, Pro and Max.
For one-off private chats
- ChatGPT Temporary chat: doesn't appear in your history, doesn't create memories, and isn't used to train OpenAI's models. OpenAI may keep it for up to 30 days for safety. Before it starts, you can choose whether it uses your existing memories and custom instructions. If you save a temporary chat, it becomes a regular chat and follows your normal settings.
- Claude Incognito chat: available on every Claude plan. Click the ghost icon in the top right when you start a new chat. Anthropic says incognito chats aren't used for training, whatever your training setting is. They aren't saved to your history, and Anthropic keeps them for 30 days by default.
Copy-paste checklist
[ ] ChatGPT: Settings > Data controls > Improve the model for everyone = OFF [ ] ChatGPT: Include your audio recordings = OFF (if shown) [ ] Claude: Settings > Privacy > model training toggle = OFF [ ] I know where Temporary chat (ChatGPT) and Incognito (Claude) are [ ] I don't press thumbs up/down on chats I want kept out of training
Common mistakes
- Thinking it deletes old chats. It doesn't. Delete chats separately if you want them gone.
- Clicking thumbs up or down. OpenAI says that if you give feedback, the whole conversation may be used for training even after you opt out. Anthropic's privacy center also says it may use your feedback to train its models.
- Forgetting the safety exception. Anthropic says chats flagged by its safety classifiers may still be used to improve its trust and safety systems.
- Assuming it's the same at work. Business, Enterprise and Edu workspaces have their own rules set by your admin. OpenAI doesn't train on them by default.
- Treating "off" as permission to paste anything. Keep passwords, card numbers and ID numbers out of AI chats either way.
