AI chatbots can be useful for writing, planning, and learning. But a chatbot is not a private diary, a bank vault, or your doctor’s secure patient portal. A simple rule protects you: share the minimum information needed for the task.
Never paste these seven things
- Passwords, PINs, or one-time verification codes. No legitimate helper needs them to explain a problem.
- Social Security, Medicare, passport, or driver’s-license numbers. These identifiers can be used for identity theft.
- Bank, credit-card, investment, or tax-account details. Remove account numbers, routing numbers, balances tied to your identity, and login information.
- Complete medical records with identifying details. If you need help organizing questions, use a general description and remove your name, birth date, record number, address, and insurance information.
- Private legal documents or security answers. Wills, contracts, signatures, case files, and answers such as a mother’s maiden name deserve professional, secure handling.
- Another person’s confidential information. A family member’s diagnosis, finances, address, or private story is not yours to upload without permission.
- Anything you would be seriously harmed to see exposed. This includes intimate photos, unreleased business information, and precise details of when your home will be empty.
Replace details with placeholders
You can often get the same help without identifying yourself. Replace real details with labels:
- “[Medication A]” instead of the full prescription label and patient number
- “a family member in their 60s” instead of a name and birth date
- “a monthly expense of about $500” instead of an account statement
- “my local clinic” instead of a full address and appointment confirmation
Know what privacy controls can—and cannot—do
Privacy settings differ by service and can change. In ChatGPT, for example, Data Controls can turn off use of new consumer conversations for model improvement. Temporary Chat does not appear in history, does not create memories, and is not used for training; OpenAI says Temporary Chats are deleted from its systems after 30 days, subject to stated safety, security, or legal exceptions.
Those controls reduce exposure, but they are not permission to paste secrets. The safest habit is still to remove sensitive details before you begin.
A safer way to ask
I need a general checklist for preparing questions about a new medication. Do not diagnose me. Do not ask for my name, birth date, medical-record number, insurance information, or account credentials.
Continue with a safer next step
Explore AI Made Practical After 50, visit the Scam & Privacy Protection hub, and learn how to spot AI-generated scams.
Sources and privacy help
- OpenAI: Data Controls FAQ
- OpenAI: Chat and file retention policies
- Federal Trade Commission: Protect your personal information
- Federal Trade Commission: Online privacy and security
Product settings and retention practices can change. Check the service’s current privacy documentation before sharing information.
