Safety checklist
What can workplace AI see?
A plain-English data visibility check for office workers before using AI with work content.
Why this matters
Your anxiety may be about access, not ability.
Many people hesitate because they do not know whether an AI tool can see emails, documents, chats, calendars, files, customer records, or only what they paste into the box.
Use today
Check the tool boundary before using the tool.
If you cannot explain what the AI can access, treat the task as sensitive and use a generic practice version instead.
The six questions to ask before using workplace AI.
Is this tool approved?
Use the AI tools your workplace has approved. If you are not sure, ask before using real work content.
What can it access?
Can it see only what you paste, or can it connect to email, files, calendar, chat, or internal systems?
Who controls it?
Is it your personal account, a company account, or a tool with admin controls and company policies?
Can people review usage?
Some workplace systems may log prompts, files, outputs, or admin events. Assume work tools may be auditable.
What should not be shared?
Keep customer, employee, financial, legal, medical, confidential, password, and private message content out unless clearly approved.
Can I use a safe version?
Replace names, numbers, clients, projects, and private facts with placeholders while learning the workflow.
Copy-paste prompt: data visibility check.
Simple visibility map.
Usually lower risk
- Public information.
- Generic examples.
- Your own non-sensitive draft text.
- Placeholders instead of real details.
Check first
- Company documents.
- Meeting notes.
- Email threads.
- Internal process details.
Do not paste casually
- Customer or employee data.
- Contracts, legal, financial, or medical information.
- Passwords, tokens, or API keys.
- Confidential strategy or private disputes.