Team AI Ground Rules

Three rules your team can agree on in 5 minutes, while the formal process catches up.

The one-page version

Rule 1 · Nothing confidential goes into a personal AI account. Use the tools our company has approved, and follow its policy.

Rule 2 · If AI wrote the first draft, say so. No shame. It tells the reviewer where the work started and what to check.

Rule 3 · A person signs off before anything leaves the team. AI can draft the email, the report or the recommendation. A named person checks the facts and owns the final version.

Team: ___ Agreed on: ___ Questions go to: ___

A 5-minute meeting script

Minute 1 · Why. "Most of us are already using AI at work. That's good. I want us to agree on how, so nobody gets caught out."

Minutes 2 to 3 · The rules. Read the 3 rules out loud. Ask: "Does anything here stop you doing your job?"

Minute 4 · Tools. "Here's what's approved for us right now: ___. If you're not sure, ask before you paste."

Minute 5 · Agree. "Can we all work this way starting today? I'll send the one-pager after this meeting."

4 questions to ask IT or your security team

• Which AI tools are approved for our team, and for what kind of information?
• What must never be pasted into any AI tool, even an approved one?
• Do approved tools keep or train on what we type?
• Who do we tell if someone pastes something they shouldn't have?

When someone breaks a rule

Keep it no-blame. The goal is to fix the gap, not punish the person.

• Thank them for telling you.
• If confidential information went into a personal account, follow your company's incident process straight away.
• Ask what made the rule hard to follow. Fix that (tool access, unclear examples, time pressure).
• Update the one-pager if the rule needs to be clearer.

Important

These rules are a starting point for your team. They don't replace your company's legal, privacy or security policy. Where your company's policy says something different, the company policy wins.