Free visual lessons · Start here
Big tech ideas. Small, clear steps.
Open a card, try an example, and build your understanding. You don’t need to memorize a wall of logos.
Who’s in charge of the AI?
AI governance means deciding how AI may be used, who is responsible, and how risks are checked. These are DCT’s introductory questions, not a complete governance framework.
01 Who gets access?
Verify people and apps, and give each only the permissions needed. Example: a workshop helper may view the class schedule without accessing private learner records.
02 What data may it use?
Agree which sources are allowed. Protect private information and honor access permissions before anything is sent to an AI service.
03 Who reviews the result?
Assign a person to check important outputs and approve consequential actions. A convincing answer can still be wrong.
04 How do we check it?
Test realistic tasks, look for errors and unfair outcomes, monitor use, and give people a way to report problems.
05 What happens when it fails?
Name an owner, define escalation steps, and decide when the system should be paused or changed.
Try it: Imagine an AI assistant replying to class inquiries. Which messages can it answer? Which should go to a person?
Explore NIST’s AI Risk Management FrameworkWhat’s under the hood of an AI assistant?
A teaching map of common components. An app may use only some of these layers.
A The place you ask
The interface collects your question and displays the response. Think of a chat window or a help form.
B The model
A language model generates responses using patterns learned during training and the context supplied to it.
C The information it can find
Retrieval can supply relevant documents to a model. Search and storage are different roles; access permissions still matter.
D The actions it can take
Connected tools can allow an app to perform actions. Limit permissions and require approval for sensitive actions.
E The checks around it
Evaluation, monitoring, privacy controls, and human review help people assess and manage the system.
Try it: Where would you put a rule that prevents the assistant from sharing private student details? Consider both data access and output review.
One task. Three ways.
Excel, Python with pandas, and SQL can all work with tabular data. Here, each row is one class signup, with a numeric age column. Python assumes the table is loaded as df; SQL assumes a table named signups. These examples explain the operation; they do not run here.
| Task | Excel | Python / pandas | SQL |
|---|---|---|---|
| Keep adults | Filter age ≥ 18 | df[df["age"] >= 18] | SELECT * FROM signups WHERE age >= 18; |
| Sort youngest first | Sort age smallest to largest | df.sort_values("age") | SELECT * FROM signups ORDER BY age ASC; |
| Average age | =AVERAGE(B2:B6) if ages are in B2:B6 | df["age"].mean() | SELECT AVG(age) FROM signups; |
Missing values, data types, and SQL dialects can affect results. Use made-up records to practice.
See the official pandas / SQL comparisonChoose your next learning step.
Curious about cloud? Start with a short lesson. Ready for practice? Preview the self-paced starter. Prefer a group? Explore the proposed live schedule.
Try a mini lessonPreview Cloud101 · $97Explore learning resourcesExplore live classes