Page 355 - CyberSecurityEssentialsEnglish

9. How does an extraction card preserve the link to the original text?
10. What is the difference among summary, comparison, synthesis, and inference?
11. What are the risks of using generated code without testing it?

12. How can AI be used in qualitative analysis without replacing human interpretation?
13. What is the structure of claim, evidence, and reasoning?

14. What are the steps of fully verifying a reference?
15. Why can an AI tool not be listed as an author?
16. What must a disclosure of a consequential use include?

17. What are the constraints on using AI in peer review?
18. What are the essential fields in an AI usage log?
19. How do reporting guidelines and pre-registration serve reproducibility?

20. Which outputs should the evidence map in the lab produce?

Practical Exercises
↻ Exercise 1: Turn a broad topic into a researchable question
Goal: Separate a topic from a problem and from a gap.
Required: A subject you are working on. Allow about 45 minutes.
Steps: Write the topic in one line. State the problem it creates and for whom. State what
is not yet known and which kind of gap it is. Write the question with its population, its
comparison where one exists, and its outcome. Define every ambiguous term

operationally. Then test the question against the five criteria of feasible, interesting,
novel, ethical, and relevant.
Output: One page with the question, the definitions, the reason for the choice, and the
scope.
Assessment criterion: The question is specific, feasible, and ethical, and the topic, the
problem, and the gap are clearly distinguished.

↻ Exercise 2: Build a repeatable search strategy
Goal: Make a search another person could run and obtain a comparable set.
Required: The question from Exercise 1 and access to one database. Allow about 60
minutes.
Steps: List the concepts and the alternative terms for each. Combine them with the
operators. Record the exact query, the database, the date, and the number of results.
Revise the query twice and record why each change was made. Then note where the

strategy remains weak.
Output: A small search log with three versions of the query, the result counts, and a
quality note.
Assessment criterion: Another person could re-run the search, and every change is
justified by the concepts rather than by the results it produced.

354