Deep dive · module 68 of 100
When you should not use AI
Recognise situations where AI is not the right choice.
After this lesson
You can name three situations in which it is better not to use AI.
In view · Choices and control
Three reasons to opt out
If you cannot assess the accuracy, choose a different approach.
Have decisions with major consequences reviewed by authorised people.
Recognise when empathy, trust, or consultation is needed.
Part 1 of 3 · Not verifiable
Your preference for less movement is on. Choose each part yourself.
When you cannot check the answer
If you cannot assess the accuracy, AI is risky.
In case of high consequences
Legal, medical, or financial decisions with a large impact.
If the task requires human contact
Empathy, trust, or negotiation remain human work.
New practice example
From explanation to a usable result.
Sometimes the correct AI result is the decision not to leave the task to a language model.
Fictional practice situationA fictitious team wants AI to independently calculate the final salary amounts from incomplete notes. An approved payroll administration and an authorised employee are available.
View a possible solution
This is how it can be done
Use the authorised employee and the existing calculation system for the amounts. AI can at most help formulate fictitious explanations or a checklist. Incomplete source data is no basis for a real financial decision.
Why this works
A model can calculate convincingly without a reliable process or complete input. The task and consequences determine whether the tool is appropriate.
Make it slightly more difficult
Choose an innocent text task from the same working environment and describe why AI can help there with a clear control.
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
A model can calculate convincingly without a reliable process or complete input. The task and consequences determine whether the tool is appropriate.
A step further
Choose an innocent text task from the same working environment and describe why AI can help there with a clear control.
A prompt to get you started
Name three situations in which it is better not to use AI, with a short reason per situation.
Apply it
A team wants to leave a difficult personnel decision entirely to a general chatbot.
Which approach best fits this situation?
- Let the chatbot make the final decision.
- Enter all personal details without necessity or appropriate agreements.
- Let authorised people decide and use at most appropriate, controlled support.
With JavaScript, you can check your approach and receive immediate feedback.
Test your insight
When is it wise not to use AI?
- If you cannot check the result or if the consequences are large.
- Never, AI is always better.
- Only if you have no time.
Extra practice
Note two tasks in your work where you would deliberately not use AI.
1. Giving legal advice. 2. A difficult conversation situation with an employee.
