Deep dive · module 93 of 100
Governance and risk management
Ensure oversight and agreements regarding AI use within the organisation.
After this lesson
You can name three governance agreements that mitigate risks.
In view · How it connects
Responsibility gets a name
Designate who may approve new applications and changes.
Determine which consequences require extra testing and control.
Establish ownership, reporting route, and follow-up of errors.
Part 1 of 3 · Who decides?
Your preference for less movement is on. Choose each part yourself.
Who decides?
Who is allowed to approve new AI tools or applications?
Assessing risks
Which applications have higher consequences and deserve extra control?
Accountability
Who is responsible if something goes wrong, and how is that reported?
New practice example
From explanation to a usable result.
Risk management links a concrete application to possible consequences, checks, and a responsible party.
Fictional practice situationA fictitious association uses the same AI tool for public texts and for organising complaints about members. The second task contains sensitive allegations.
View a possible solution
This is how it can be done
Create two separate usage agreements. For public texts, check source facts before publication. For complaints, extra assessment of access, data, reliability, and human handling is required before the application can be used.
Why this works
The same technique can have different consequences. A general tick of "AI permitted" does not answer the practical risk questions.
Make it slightly more difficult
Note who can report an error, who decides on stopping, and what evidence is required before you reuse the application.
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 same technique can have different consequences. A general tick of "AI permitted" does not answer the practical risk questions.
A step further
Note who can report an error, who decides on stopping, and what evidence is required before you reuse the application.
A prompt to get you started
Name three governance agreements an organisation can make to manage AI risks.
Apply it
A new AI tool will handle customer queries. No one is appointed to assess risks or follow up on incidents.
Which approach best fits this situation?
- Define approval, responsibility, notification, and evaluation before implementation.
- Let the vendor take over all responsibility without agreements.
- Wait for an incident before making agreements.
With JavaScript, you can check your approach and receive immediate feedback.
Test your insight
What is the purpose of AI governance?
- Maintaining oversight and managing risks without blocking innovation.
- Banning all AI.
- Only controlling costs.
Extra practice
Write one concrete governance agreement that would suit your organisation.
New AI tools for customer contact must first be approved by the team leader and the privacy officer.
