Your AI Safety Checklist

Learning Objectives

• Have a ready-to-use checklist for AI safety and privacy
• Know exactly which settings to change and which habits to build
• Be able to reference this lesson as an ongoing resource

Your AI Safety Checklist

This lesson is designed to be bookmarked and revisited. Work through each section at your own pace — you do not have to do everything at once. The most important items are marked with a star.

Privacy Settings (Do These First)

These are the highest-impact actions. Each one takes under 2 minutes.

ChatGPT:

• Turn off "Improve the model for everyone" in Settings → Data Controls
• Review and clear Memory if it contains personal information
• Use Temporary Chat for anything sensitive
• Review and delete old conversations you no longer need

Claude:

• Review conversation history and delete sensitive conversations
• Note: Anthropic does not train on free-tier conversations by default

Google Gemini:

• Turn off Gemini Apps Activity in Google Account settings
• Review and delete saved conversations
• Check that SafeSearch is enabled if you have children

Microsoft Copilot:

• Review Activity history in Microsoft Account → Privacy dashboard
• Check enterprise data handling policies with your IT team if applicable

General (all platforms):

• Use a unique, strong password for each AI platform
• Enable two-factor authentication where available
• Review connected apps and revoke access to any you no longer use

Data Sharing Habits

Build these into your daily AI use:

Before typing any prompt, ask:

• Would I be comfortable if this became public?
• Does this contain anyone else's personal information?
• Does this contain business-confidential information?
• Could I rephrase this to remove identifying details?

Information you should never share with AI:

• Passwords, API keys, or authentication credentials
• Social Security numbers, national ID numbers, or financial account details
• Other people's personal information without their consent
• Classified, legally privileged, or NDA-protected information
• Medical records or health information (yours or others')

Information to share cautiously (anonymize first):

• Work scenarios involving specific people (remove names and identifying details)
• Business strategies (generalize the industry and company details)
• Personal situations you want advice on (remove names and locations)

Verification Habits

Protect yourself from AI-generated misinformation:

When reading content:

• Check the source — who published this and what are their standards?
• Cross-reference dramatic claims with at least one other reputable source
• Be skeptical of content that triggers a strong emotional reaction
• Look for specific, verifiable details rather than vague claims

When viewing images or video:

• Check hands, text in images, and background consistency
• Use reverse image search for suspicious photos
• Watch for lip-sync issues and face-edge artifacts in video
• Ask: does it make sense that this person would say or do this?

Before sharing anything:

• Pause. Take 30 seconds before sharing content that provoked a strong reaction.
• Verify. Can you confirm the key claim from a second source?
• Consider. What happens if this turns out to be false and you shared it?

Workplace AI Safety

If you use AI at work:

• Know your company's AI policy — if there is no policy, assume conservative data handling
• Never paste confidential business data into free-tier AI tools
• Disclose AI assistance when your company or client requires it
• Use enterprise-grade AI tools (with business data protections) for sensitive work
• Keep records of how you use AI — this protects you if questions arise

Family AI Safety

If you have children who use AI:

• Review the privacy settings on any AI tool your child uses
• Talk about what is and is not safe to share with AI (names, school, location, photos)
• Set up parental controls on devices — see our AI for Parents playbook for detailed guidance
• Teach the verification habit: "How do you know that is true?"
• Monitor for emotional dependence on AI chatbots

Red Flags to Watch For

Stop and reconsider if:

• An AI tool asks for access to your contacts, files, or accounts beyond what is needed for its function
• A "free" AI tool has no clear business model — your data may be the product
• An AI service makes promises about privacy that seem too good ("we never store anything")
• You receive AI-generated content that seems designed to make you angry, fearful, or excited enough to share without thinking
• Someone asks you to bypass your company's AI policy "just this once"

Monthly Maintenance

Set a recurring monthly reminder to:

1. Review and delete old AI conversations across all platforms
2. Check privacy settings — platforms sometimes reset them after updates
3. Review any new AI tools you have started using and check their data policies
4. Update passwords if any AI platform has reported a security incident

Key Takeaways

• Privacy settings are the single highest-impact action — do them first
• Build the "pause before typing" habit for every AI interaction
• Never share credentials, others' personal data, or NDA-protected information
• Verify before sharing content that triggers strong emotions
• Set a monthly reminder to review your AI privacy and security hygiene
• Bookmark this checklist and revisit it when you adopt new AI tools