Deep dive · module 28 of 100
Creating AI policy
Simple agreements in your team about what is and is not allowed.
After this lesson
You write three concrete agreements for AI use in a team.
In view · How it connects
AI policy on one page
Write down what your team is allowed to use AI for.
Make it clear which information must not be entered.
Specify who checks and where questions or incidents should be directed.
Part 1 of 3 · Permitted tasks
Your preference for less movement is on. Choose each part yourself.
Minimum elements
Permitted tools, what you may input, when to check, who is responsible. Also describe who approves an application, when a process must stop, and how an error is reported. A list of permitted tools alone does not yet say which data or actions are permitted within them. Link rules to recognisable tasks.
Keep it practical
Short agreements that everyone understands work better than long documents. For example, specify what an employee does in the event of a missing price in a customer email: do not supplement or send, but consult the product owner. Concrete examples make an agreement actionable. This does not replace an assessment of a real application.
Updating
Evaluate after a few weeks what works and adjust. Assign an owner, version date, and planned evaluation moment to the policy. Use feedback from real, permitted trials to improve agreements. Report changes to the team and prevent an outdated version from continuing to circulate as current instruction.
New practice example
From explanation to a usable result.
A useful AI agreement states who may do what, with which data and which control.
Fictional practice situationFiction: a small team wants to use AI for general newsletters. It only uses approved public information for this purpose.
View a possible solution
This is how it can be done
Agreement: one employee creates a draft with approved information; a second employee checks facts, tone, and dispatch. New applications are first discussed with the agreed responsible person.
Why this works
The rule links actions to roles and control moments. This lesson example does not replace a full legal or organisational assessment.
Make it slightly more difficult
Describe how the agreement is adjusted if the tool or the type of input data changes.
Extra practice, at your own pace. You do not need to submit anything. This does not count towards an automatic assessment.
Your workplace · learning by doing
Make something. See what changes.
Checking the AI connection… You can already view the written example.
Present your assignment to Ende.
Read the answer, change one thing, and compare.
Previously held
New version
You remain in control
What do you want to improve in your next attempt?
On this device only. AI answers remain only in this page memory. This is self-review, not a grade.
Why this approach works
The rule links actions to roles and control moments. This lesson example does not replace a full legal or organisational assessment.
A step further
Describe how the agreement is adjusted if the tool or the type of input data changes.
Try & discover · no obligation
Choose which data to omit
View the fictitious data. Omit sensitive practice fields and retain only what the task requires.
Fictional practice data. Names, addresses and codes are examples. Your task is to prepare a general text; sensitive file details and secrets are not required for this.
What would you include as input?
Example key: VOORBEELD-SLEUTEL-GEEN-ECHTE Confidential team plan: Non-public practice appointments for next quarter. Task: Create general team agreements regarding permitted input and human control.
Still visible in this exercise: example key, confidential team plan. Are those fields necessary for the general task?
Omitting or masking is not proof of anonymity. This fixed exercise does not check real personal data, files, or policies.
This trial works locally with fictitious data and does not count towards your lesson completion. It starts anew with a new lesson or reload.
A prompt to get you started
Help me draft three short AI agreements for a small team. Topics: data, control, responsibility. Keep it simple. Use only the provided fictitious facts. Make missing information visible and indicate what a human must check before use.
Apply it
A team rule only says 'Use AI responsibly'. An employee does not know what to do if a price is missing in the source.
Which addition makes the agreement actionable?
- Primarily use your common sense.
- Always let the tool estimate a price.
- Do not enter a price and do not send yet; ask the product owner for confirmation and use the agreed reporting point for errors.
With JavaScript, you can check your approach and receive immediate feedback.
Test your insight
Why is a AI policy useful?
- It replaces all control.
- It provides clear agreements about use and responsibility.
- It is only for large companies.
Extra practice
Write three concrete agreements that you would put in a AI policy for your team. Work with a fictitious situation and name one concrete control point before you would use the result.
1. No personal data in public tools. 2. Always check before sending. 3. If in doubt, ask the team lead. I compare the outcome with the confirmed input, leave unknown data blank, and adjust the result before use.
