Review · module 104 of 155
Review: working safely and responsibly
Bring data protection, fairness, transparency and agreements together in one way of working.
After this lesson
You choose which data you may use, recognise when AI is not appropriate and know who is responsible for a decision.
In view · How it connects
Three questions before you use AI
Only give what the task needs, in a permitted environment.
When the consequences for people are serious, a responsible person decides, not the tool.
Policy and governance say what is allowed, who checks and where you report errors.
Part 1 of 3 · Which data?
Your preference for less movement is on. Choose each part yourself.
Share only what is necessary and permitted
Look at your input before you put it into an AI tool. Personal data, confidential information and secrets such as passwords do not simply belong there (lessons 5 and 23). Use the tool and environment your organisation has approved, and only give the data the task really needs. Leaving out a name does not always make a file anonymous: a job title, date and place together can make someone identifiable. Practise with fictional data and stop if you are in doubt about permission.
Know when AI is not the right choice
Some tasks require a person with responsibility: decisions about staff, health, rights or people’s money (lesson 68). AI can then at most provide support, with clear checks. Watch out for bias: a model can repeat stereotypes or treat groups differently (lesson 83). Test with a range of different examples and look at the outcome per group. Be open about AI use where it matters, so that readers know how a text or decision came about (lesson 84).
Agreements make it concrete
An AI policy says which tools and data are permitted, who checks and where you report errors (lesson 28). Governance links each application to an owner, a risk assessment and follow-up (lesson 93). A team strategy determines what you want to use AI for and how people learn (lesson 92). Among other things, the European AI Act requires organisations that use AI to take measures to support the AI literacy of their staff. For specific questions, read the current official text or seek advice.
New practice example
From explanation to a usable result.
Use only data that is necessary and permitted, choose an approved tool and let people decide where the consequences are serious.
Fictional practice situationFictional: an association receives a complaint about a sports kit that was delivered late. The message contains the name, the address, the membership number and a note that the complainant’s child has an injury.
View a possible solution
This is how it can be done
For a general reply letter about delayed deliveries, only the complaint is needed: the kit arrived late. I leave out the name, address, membership number and the injury. The person responsible at the association checks the letter and sends it from the association’s own membership system.
Why this works
Sharing less data reduces the risk without harming the task. Information about health is extra sensitive and not needed here.
Make it slightly more difficult
Think of a task where a piece of personal data is needed. What agreement or tool must be in place before you use AI?
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
Sharing less data reduces the risk without harming the task. Information about health is extra sensitive and not needed here.
A step further
Think of a task where a piece of personal data is needed. What agreement or tool must be in place before you use AI?
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 complaint file with practice values. Your task is to prepare a general reply letter about a late delivery. Identification and health data are not needed for that.
What would you include as input?
Name: Fictional customer Lien Voorbeeld National register number: 00.00.00-000.00 (practice value) Medical reason: Fictional: could not collect the delivery because of a hospital stay. Complaint: The delivery arrived two days later than promised.
Still visible in this exercise: name, national register number, medical reason. 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 prepare a fictional complaint file for a general reply letter. The file contains these types of data: name, national register number, phone number, a medical reason and the complaint itself: a delivery arrived two days late. For each type of data, say whether it is needed for a general reply letter and why. Then write a short general letter without personal data.
Apply it
A team member wants to let a free online AI tool choose which applicants are invited, by pasting all the CVs into it.
Which approach best fits this situation?
- Carry on if the tool gives a clear ranking.
- Only remove the names from the CVs and then let the tool decide.
- Stop this approach: only use a permitted environment for CVs and let a person decide, with AI at most in a supporting role within the agreements.
With JavaScript, you can check your approach and receive immediate feedback.
Test your insight
You want to have a general reply letter drafted. Which input is the most responsible?
- The complete file, so the AI has all the context.
- Only the general description of the complaint, in a permitted tool.
- The file without the name, because then it is anonymous.
Extra practice
For a task from your own work, make a short list: which data is needed, which do you leave out, which tool is permitted and who decides on the result?
Needed: the complaint, a delivery that arrived two days late. Leave out: name, national register number, phone number and the medical reason; they are not needed for a general text. Tool: only the approved business environment. Decision: customer service approves the letter and only adds personal data afterwards, in its own customer system.
