AI at Work Safety Checklist

Before you start

• I know which AI tool my employer allows, and I'm signed in with my work account for work.
• I've checked the tool's data settings: training on my chats, memory and chat history.
• I know which apps it is connected to (email, calendar, files) and removed what it doesn't need.

What never goes in

• No passwords, account numbers or security codes.
• No customer or employee personal details (names, emails, phone, addresses).
• No health, HR or legal case details.
• No confidential financials, contracts or unreleased plans.
• No screenshots or recordings that show other people's details or other open tabs.
• When in doubt, anonymize: “Client A,” “$X.”

Before you let AI act

• I said which sources it may use and what it may not touch.
• It asks me before it sends, posts, buys, deletes or changes anything.
• I review the result, and what it did, before anyone else sees it.
• If it acts strangely after reading a page or file, I stop and tell IT.

Before you use the answer

• I checked every number, date, name and link against a real source.
• I opened the sources it cited and they say what it claims.
• I asked “What might be wrong or missing?”
• I compared any AI meeting summary with my own notes.
• It sounds like me and includes details only I would know.
• I'm comfortable being accountable for it.

Build the habit

• I picked one workflow to use daily for two weeks.
• I'm tracking the minutes it saves.
• I re-check my settings when the tool adds a big new feature.
• I shared one tip with a colleague.