What You Should Never Paste Into an AI Chatbot

The safest rule is simple: don't paste passwords, authentication codes, private keys, highly sensitive personal records, or somebody else's confidential information into a general-purpose AI chatbot unless you have specifically verified that the service and account are approved for that data.

AI services have different privacy controls, retention policies, account types, and data practices. Don't assume they all work alike.

Never paste passwords or authentication secrets

An AI assistant doesn't need your actual password to explain how password managers work.

It doesn't need your two-factor authentication code to explain two-factor authentication.

And it doesn't need an API secret or recovery code to help you understand an error message.

Replace secrets with obvious placeholders such as YOUR-PASSWORD-HERE.

Be careful with documents

Before uploading a document, ask what is actually inside it.

A harmless-looking PDF could contain an address, account number, customer information, employee records, signatures, internal business information, or metadata you didn't intend to share.

Remove unnecessary sensitive information first.

Privacy controls vary by service

This is where generic advice becomes dangerous.

For example, OpenAI says Temporary Chats are not used to improve its models while they remain temporary, don't normally appear in chat history, and may still be retained for up to 30 days for safety purposes.

Google separately documents what information users provide to Gemini Apps, including prompts, files, videos, screens, photos, imported chats, and shared page content.

Those are service-specific rules, not universal promises about AI.

Work information needs another check

If your employer gives you an approved AI service, follow your organization's policy.

A consumer account and an organization-managed account can have different terms and controls.

Don't paste customer information or confidential work material into a personal AI account merely because the chatbot makes the task convenient.

AI also makes mistakes

Privacy isn't the only issue.

An AI system can confidently produce an incorrect interpretation of the information you give it. Treat consequential answers as material to verify, not unquestionable authority.

A useful habit

Before pressing Send, ask: Would I be comfortable if this exact information were stored outside my device under the service's stated policy?

If the answer is no—or you don't know what the policy says—remove the sensitive information first.

Sources and Further Reading

OpenAI Help Center: Temporary Chat FAQ — https://help.openai.com/en/articles/8914046

Google Gemini Apps Help: Gemini Apps Privacy Hub — https://support.google.com/gemini/answer/13594961?hl=en