Turn on the privacy settings in your AI apps
ChatGPT
Settings, then Data Controls, then turn off Improve the model for everyone. For one-off private questions use Temporary Chat: it isn't saved to history, doesn't create memories and isn't used for training, though OpenAI keeps it for up to 30 days for safety. Memory is under Settings, then Personalization.
Microsoft Copilot
Select your profile icon, then your name, then Privacy, and turn off both training switches: one for text conversations and one for voice (currently labelled Training on conversation activity and Training on voice conversations; Microsoft updated Copilot in 2026, so wording may differ). Memory is under your profile, then Memory.
Google Gemini
In Gemini, open Settings, then Activity, and turn off Keep Activity (it used to be called Gemini Apps Activity). You can also use a Temporary Chat, which isn't used for training and is kept for up to 72 hours.
Claude
Settings, then Privacy, then turn off Help improve Claude. With it on, Anthropic may keep chats for up to 5 years for training; with it off, the standard retention is 30 days.
Meta AI (Facebook, Instagram, WhatsApp, Messenger)
In the US there is no switch to stop it. Since December 16, 2025, Meta uses your Meta AI chats to personalise ads, and public posts can be used for training. The only real control is not telling Meta AI anything personal.
Apple Intelligence
Most requests run on the device, and larger ones use Apple's Private Cloud Compute, which Apple says doesn't keep your data. The optional ChatGPT extension is off by default. Check it under Settings, then Apple Intelligence & Siri, then ChatGPT, and keep Confirm ChatGPT Requests on so you're asked each time.
Check your understanding
Which of these apps has no US opt-out from using your chats: ChatGPT, Claude or Meta AI?
Meta AI. ChatGPT, Claude, Gemini and Copilot all have a switch.
What is a 'temporary chat'?
A conversation that isn't saved to your history or used for training, although it's kept briefly for safety.