Page 298 - CyberSecurityEssentialsEnglish

Chapter Objectives
Understand the roles AI can play across the research cycle, and the limits that should not be
crossed.

Turn a broad topic into a problem, a gap, a research question, and operational definitions
that can be studied.
Design a search strategy covering databases, keywords, synonyms, and full documentation

of the queries.
Screen sources and appraise relevance, quality, and risk of bias, while recording reasons for
inclusion and exclusion.

Read papers and extract evidence into a matrix that preserves the identity of the source, the
location of the information, and its limits.

Move from summarizing individual studies to a synthesis that explains agreement,
difference, and uncertainty.
Use AI cautiously in quantitative and qualitative analysis and in writing and reviewing code.
Write evidence reviews, research abstracts, and policy briefs that tie every claim to an

appropriate source.
Verify references, quotations, identifiers, and publication status, and detect fabricated or

unsuitable references.
Apply rules of integrity, authorship, disclosure, and confidentiality in research, peer review,
and publication.

Create a research record documenting the tools, versions, prompts, decisions, files, and
revisions.
Produce a documented evidence map starting from a research question in the practical lab.

FIGURE 7-1: THE AI-SUPPORTED RESEARCH CYCLE WITH HUMAN CHECKPOINTS
What to notice: A tool can assist at almost every stage, but the question, the method, the
checking of sources, the interpretation, and the responsibility do not transfer to it.
Source: Prepared by the author, drawing on guidance for generative AI in research and on

research integrity standards [1][11].

297