How to Check Your AI Privacy Settings (All Major Platforms)

This guide walks you through the exact steps to check and disable AI training on the major platforms. No screenshots included yet, but the instructions are current as of November 2025.

Related: Stop Worrying About AI Stealing Your Data covers when you actually need business tier plans vs. consumer plans with privacy settings configured.

---

Claude (Anthropic)

Important context: Anthropic changed their policy in late 2025. Consumer accounts (Free and Pro) now have training enabled by default. You must opt out manually.

How to check:

1. Go to claude.ai and sign in
2. Click your profile icon in the bottom-left corner
3. Select Settings
4. Click the Privacy tab at the top
5. Scroll down to Privacy settings
6. Look for "Help improve Claude"

What to look for:

• Toggle ON (blue): Your chats ARE being used for training
• Toggle OFF (gray): Your chats are NOT being used for training

To disable:

Click the toggle to turn it OFF. The setting saves automatically.

What you should know:

• The setting applies to new conversations from the moment you change it
• If you revisit old conversations after the policy change, those can be included
• Disabling this extends data retention from 5 years back to 30 days
• Enterprise and Team accounts are exempt by default

---

ChatGPT (OpenAI)

How to check:

1. Go to chatgpt.com and sign in
2. Click your profile icon in the bottom-left corner
3. Select Settings
4. Click Data controls
5. Look for "Improve the model for everyone"

What to look for:

• Toggle ON: Your conversations ARE being used for training
• Toggle OFF: Your conversations are NOT being used for training

To disable:

Click the toggle to turn it OFF.

What you should know:

• This applies to your entire account across all devices
• The opt-out is not retroactive for past conversations
• Important caveat: If you give thumbs-up or thumbs-down feedback on a response, that entire conversation may still be used for training, even if you've opted out
• ChatGPT Team, Enterprise, and API users are not included in training by default
• Temporary Chats are deleted after 30 days and not used for training

---

Google Gemini

Important context: As of September 2025, Google uses your uploads (files, photos, videos, screenshots) for training by default.

How to check:

1. Go to gemini.google.com and sign in
2. Click the Activity icon (clock with arrow) in the left sidebar
3. Look for "Gemini Apps Activity" (may be called "Keep Activity")

What to look for:

• Activity ON: Your chats and uploads ARE being used for improvement
• Activity OFF: Your data is NOT being saved or used

To disable:

1. Click "Turn off Gemini Apps activity"
2. Confirm by clicking again in the dropdown

What you should know:

• Turning this off means you lose conversation history (you can't refer back to previous chats)
• Responses may become more generic without personalization
• Chats already reviewed are kept separately for up to 3 years, even if you delete your activity
• Google Workspace accounts managed by organizations are exempt by default
• This is separate from Gmail's "smart features" which scan emails for spam filtering and suggestions (that's not AI model training)

---

Microsoft Copilot

Microsoft Copilot has different privacy controls depending on which version you're using.

For Personal Accounts (Copilot Web/Desktop):

How to check:

1. Go to copilot.microsoft.com and sign in
2. Click your Profile icon in the top-right corner
3. Select Privacy settings or click your account name
4. Look for "Model training"

To disable:

Toggle "Model training" to OFF.

What you should know:

• You can also disable Personalization which uses your chats and Bing activity
• Opting out of training doesn't exclude your data from being used for "general product improvements" or advertising purposes per Microsoft's privacy statement
• You can opt out of training and still keep personalization on

For Microsoft 365 Copilot (Work/School):

Your data is NOT used for training foundation models by default. Prompts, responses, and file contents aren't used to train the underlying AI.

Microsoft may use optional feedback to improve Copilot, but not to train the foundation models.

Gaming Copilot (Xbox Game Bar):

This has separate toggles for voice and text model training. Check Copilot privacy settings in the Game Bar if you use this feature.

---

Perplexity

How to check:

1. Go to perplexity.ai and sign in
2. Click your account name or profile icon
3. Scroll down to the Account section
4. Look for "AI Data Retention"

What to look for:

• Toggle ON (blue): Your searches ARE being used for training
• Toggle OFF: Your searches are NOT being used for training

To disable:

Toggle "AI Data Retention" to OFF.

What you should know:

• AI Data Retention is ON by default for all plans except Enterprise
• Opting out only applies going forward, not retroactively
• Perplexity does not delete training data collected before you opted out
• If you use Perplexity without logging in, your data is collected by default
• You can use Incognito mode for searches that are never stored
• You can also clear or disable Memory in Settings > Personalize > Memory
• Enterprise accounts are never included in training

---

Grok (xAI / X)

Important context: X (formerly Twitter) enables data sharing with Grok by default, including your posts and interactions on the platform.

How to check (Web):

1. Go to x.com and sign in
2. Click More in the left sidebar
3. Select Settings and privacy
4. Click Privacy and safety
5. Under "Data sharing and personalization," click Grok
6. Look for the training toggle

How to check (Mobile App):

1. Tap your profile icon
2. Tap Settings & Support > Settings and privacy
3. Tap Privacy and Safety
4. Tap Grok
5. Look for the training toggle

What to look for:

The setting says: "Allow your posts as well as your interactions, inputs, and results with Grok to be used for training and fine-tuning"

To disable:

Turn OFF the toggle. You can also click "Delete conversation history" to remove past Grok conversations.

What you should know:

• Your public X posts and interactions are used for training by default
• Making your account private prevents your posts from being used
• Private Chat mode (where available) is not used for training
• If you give thumbs up/down feedback, that conversation may still be used for training even if you opted out
• Opting out doesn't delete your data from X, it just stops AI training use

---

Quick Reference

| Platform | Default Setting | Where to Find It | Business Tier Exempt? | |----------|----------------|------------------|----------------------| | Claude | Training ON | Settings > Privacy > Help improve Claude | Yes | | ChatGPT | Training ON | Settings > Data controls > Improve model | Yes | | Gemini | Activity ON | Activity icon > Gemini Apps Activity | Yes (Workspace) | | Copilot | Training ON | Profile > Privacy settings > Model training | Yes (M365) | | Perplexity | Retention ON | Account > AI Data Retention | Yes (Enterprise) | | Grok | Training ON | Settings > Privacy > Grok | Private Chat only |

---

The Pattern

Every major platform follows the same pattern:

1. Consumer accounts have training enabled by default
2. You must manually opt out
3. Business/Enterprise tiers are exempt by default
4. Opting out is not retroactive
5. Giving feedback (thumbs up/down) often overrides your opt-out

If you're using AI for client work on a consumer plan, check these settings today. It takes 5 minutes and it matters.

---

Last verified: November 2025. These interfaces change frequently. If something doesn't match, look for "Privacy," "Data," or "Training" in your settings.