5 Privacy Settings You Should Change on Your AI Tools Right Now

AI tools like ChatGPT, Microsoft Copilot, Google Gemini, and smart home assistants have become part of daily life for millions of people. They can draft emails, summarize documents, control lights, or just answer questions. But convenience has a trade-off: many of these services are set up to collect and retain more data than most users realize. If you haven’t looked through your privacy settings lately, now is the time.

What Happened

In June 2026, The Wall Street Journal published a report titled “How to Maintain Our Privacy in the AI Age,” highlighting the default data-sharing practices of popular AI platforms. The article noted that many tools automatically save chat histories, use your conversations to train future models, and request permissions that go well beyond what is strictly needed for the feature you are using. While AI companies have added privacy controls over the past couple of years, they are often buried in menus or turned off by default.

Why It Matters

If you treat an AI assistant the same way you treat a search engine, you may be sharing far more than you intend. Chat logs can contain sensitive personal information—names, addresses, medical concerns, workplace details, or financial thoughts. That data may be stored on company servers for months or even used to improve the very model you are talking to. Unlike a private web search, an AI conversation is not necessarily anonymous. And once data is used for training, it can be difficult to remove completely.

The Electronic Frontier Foundation (EFF) has long advised that regular privacy audits are a good habit for any online service. AI tools are no exception. The good news is that the steps you need to take are straightforward and take only a few minutes.

What Readers Can Do

Below are five quick changes you can make today. Not all options are available on every platform, but most major services now support at least some of them.

1. Turn off chat history saving

ChatGPT, Copilot, and Gemini all allow you to prevent your conversations from being stored. When history is off, chats may still be kept temporarily for moderation or to improve the system, but they are not saved to your account long-term.

  • ChatGPT: Go to Settings → Data controls → disable “Chat history & training.”
  • Copilot: In the privacy dashboard (sign in at account.microsoft.com), find “Copilot settings” and toggle off “Save your Copilot chats.”
  • Gemini: Visit myactivity.google.com, find “Bard / Gemini activity,” and set the auto-delete to 3 months or turn it off entirely.

2. Opt out of data training

Most generative AI models can use your input to fine-tune future versions. You can often opt out, though the setting may be separate from chat history.

  • OpenAI: In the same Data controls menu, toggle off “Improve the model for everyone.”
  • Google: In Gemini settings, turn off “Gemini Apps activity” (this also prevents your data from being used for training).
  • Meta, Anthropic, and others typically offer similar controls in their privacy or account settings pages.

3. Restrict location permissions on AI-enabled apps

Many apps that include AI features request location access even when it is not needed for the core function. For example, a general-purpose chatbot does not need to know your exact GPS coordinates.

  • On your phone, go to Settings → Privacy → Location Services and review which apps have access. Set AI apps to “While Using” or “Never.”
  • For smart speakers and displays, check whether location services are enabled for weather or traffic features. If you do not use those, turn location off entirely.

4. Enable two-factor authentication and use app-specific passwords

This is a basic security step that also protects your AI account from being hijacked. Even if your data is safe on the provider’s side, a leaked password could expose your chat history.

  • Go to the account security settings of each AI service and enable two-factor authentication (via SMS or an authenticator app).
  • If the service supports app-specific passwords (for example, when connecting an API), use them instead of your main password.

5. Schedule a monthly privacy review

Settings change. New features are added. A quick monthly check—maybe on the first day of the month—keeps you in control. Look for any permissions or data-sharing flags you may have accidentally enabled, and review the auto-delete periods for stored conversations.

Bonus: Consider using browser extensions like Privacy Badger or uBlock Origin to block third-party trackers that some AI tools embed on their websites. These can reduce the amount of profiling data sent to advertisers.

Sources

  • “How to Maintain Our Privacy in the AI Age,” The Wall Street Journal, June 23, 2026.
  • Electronic Frontier Foundation, “How to Get the Most Out of AI Tools While Protecting Your Privacy,” accessed July 2026.
  • OpenAI privacy center, Google safety center, Microsoft privacy dashboard (current settings pages).

None of these steps will make you completely invisible online, but they will give you a much better starting point. The tools are yours to use—you should decide what they learn about you.