AI Governance Just Changed: 3 Privacy Settings You Need to Update Today

New rules around artificial intelligence are coming into effect, and they affect how companies like OpenAI, Google, and Microsoft handle your personal data. If you use ChatGPT, Copilot, or Google Gemini, the changes are worth paying attention to—not because they’re dramatic overnight shifts, but because they shift the ground beneath how your information is collected, stored, and shared.

Here’s what actually changed, why it matters for your privacy, and three specific settings you should check this week.

What happened: AI governance lands on privacy’s desk

In June 2026, a wave of AI governance frameworks matured into enforceable guidelines. The International Association of Privacy Professionals (IAPP) has been tracking these developments closely, and the headline is that regulators in several regions—including the EU, parts of the US, and the UK—are now requiring AI providers to be more transparent about training data, user inputs, and model outputs.

The IAPP’s analysis, published under the title When AI governance lands on privacy’s desk, notes that these rules don’t just apply to developers. They create obligations for service providers to give users clearer options about how their conversations and personal data are used for model improvement, testing, and profiling.

What this means in practice: the fine print you agreed to last year may no longer be the full picture. Some companies have already updated their privacy policies, and others are quietly rolling out new consent screens and dashboard controls. The catch is that most users won’t notice unless they go looking.

Why it matters for ChatGPT, Copilot, and Gemini

The core issue is that many AI tools, by default, save your conversations and use them to train or fine‑tune their models. Under the new governance rules, companies must make opting out easier and clearer, but the defaults often remain unchanged. That means if you don’t adjust your settings, your data may still be used for internal training, even when the company is technically compliant.

For example:

  • ChatGPT (OpenAI) now offers a more prominent “Turn off training” toggle, but it’s still buried under account settings.
  • Microsoft Copilot lets you turn off history and training data collection, but the setting defaults to on.
  • Google Gemini (formerly Bard) gives you a dashboard to review and delete your activity, but the auto‑save option is enabled by default.

Knowing where these controls are is more important than ever because the new governance rules give you the right to object—but only if you know how to exercise it.

3 privacy settings to update today

Take 10 minutes to check these three settings. They apply across most major AI assistants.

1. Turn off training on your data

Most AI platforms have a toggle or checkbox that stops your conversations from being used to improve the model.

  • ChatGPT: Go to Settings → Data Controls → turn off Improve the model for everyone.
  • Copilot (web or app): Go to Privacy & Safety and disable Allow Microsoft to use your chat history for training.
  • Gemini: In your Google Account, navigate to Data & Privacy → AI & Gemini → disable Gemini Apps activity.

Once you do this, your future chats won’t be fed back into the training pipeline. (Historical chats may still have been used, but you can also delete them—see next step.)

2. Delete your existing chat history

Even if you turn off future training, any conversations already saved might have been used. Deleting them removes that data from the company’s active storage.

  • ChatGPT: In the main chat list, you can delete individual conversations or clear all history in Settings.
  • Copilot: Use the Microsoft account privacy dashboard to delete your Copilot activity.
  • Gemini: The same Gemini Apps activity page in your Google Account allows bulk deletion of past interactions.

Note: some companies retain anonymized or aggregated data for longer periods, but removing the raw chat logs limits what can be associated with you.

3. Review how your data is shared with third‑party services

Many AI tools are integrated into other apps—like Copilot in Office 365 or Gemini in Gmail. The new governance rules require clearer disclosures about when your data leaves the main service. Check if your AI tool is allowed to share your data with plugins, connectors, or “workspaces.” If you don’t need those integrations, disable them.

  • ChatGPT: Settings → Connected apps → revoke any apps you don’t actively use.
  • Copilot: Microsoft 365 → Copilot settings → disable plugins that don’t add value.
  • Gemini: Google Workspace → Gemini add‑ons → disable extensions you haven’t reviewed.

How to stay informed without getting overwhelmed

AI governance is moving fast, but you don’t need to read every new law. A few practical steps:

  • Set a calendar reminder once every three months to revisit privacy settings on your most used AI tools. Companies often quietly change defaults during updates.
  • Follow the IAPP blog (their When AI governance lands on privacy’s desk post is a good starting point) for high‑level summaries without legal jargon.
  • Watch for emails from the services you use that mention “updated privacy policy” or “data processing changes”—these are your best signal that something has changed.

The bottom line for consumers

The new governance rules are a net positive: they force companies to be more upfront about how they use your data. But they don’t automatically protect you—you still need to exercise the controls they expose. The three settings above are the most impactful changes you can make in under 15 minutes. You don’t need to stop using these tools; you just need to make sure they’re working for you, not the other way around.

Sources

  • IAPP (International Association of Privacy Professionals), “When AI governance lands on privacy’s desk,” June 2026.
  • OpenAI privacy policy and ChatGPT data controls documentation.
  • Microsoft Copilot privacy settings help page.
  • Google Gemini / Google Account data & privacy controls.