
Treat an AI chat like an email to a company you trust but do not control. Leave out passwords, ID numbers, health and financial details, and other people’s private information.
Never paste these
- Passwords, API keys and security codes.
- Social Security, passport or driver’s license numbers.
- Full bank or card numbers.
- Medical records or other people’s health details.
- Confidential client or employer data your company has not approved.
Check your settings
- Data controls: whether your chats can be used to improve models.
- Memory: what the assistant remembers about you, and how to delete it.
- Temporary or incognito chats for one-off sensitive questions.
- Connected apps: what each integration can read or do.
Anonymize instead
You can still get help with sensitive work: replace names with roles (“Client A”), round numbers and remove identifying details. The advice is usually just as good.

AI Beginner Bundle
Five beginner guides that take you from AI-curious to AI-confident.
$0+ · Look insideAt work, follow the policy
Many companies approve specific tools and plans with stronger data protections. If yours has no policy yet, ask for one; it protects you as much as the company.
Watch what agents can do
When an assistant can read your email or files, hidden instructions in a web page or document could try to misuse that access. Keep approvals on for sending, buying and deleting.
Draft a simple team policy with our AI Use Policy Generator.
Put it into practice
Open the free AI Use Policy Generator: a clear AI policy for your team in minutes. This article is part of The AI Playbook for Small Businesses and Freelancers.
Sources & further reading
Product facts were checked against official sources on 16 October 2026. AI tools change quickly, so check the provider’s current documentation for anything you rely on. Our editorial policy.


