Page 276 - CyberSecurityEssentialsEnglish

✕ Common misconception Correction

Define the permitted uses with examples,
Adding the sentence “the use of AI is
tie the rule to the learning outcome of
prohibited” to a course outline and
each task, and state how disclosure should
considering the matter settled.
be made.

⚠ Warning: never accuse on the basis of a detector alone
Do not accuse a learner of relying on AI on the basis of a single detection tool or a stylistic
impression. Follow a fair process based on several kinds of evidence, and allow
explanation and review.

● Remember: Section summary

• Key idea: Integrity means the work reflects its author’s learning and that help is
declared and permitted.
• Practical decision: Set the rules of each task and redesign assessment to gather several
kinds of evidence.

• Mistake to avoid: Relying on a blanket prohibition or on an automated detector as
decisive proof.

6.7.1 Extended Case Study: Redesigning a Take-Home Assignment
This is a composite teaching case inspired by patterns common in educational institutions,

not a report about any particular one. The original task in a general university course asked
students to write a 1,200-word essay on the effect of AI on the labor market. Before
generative tools became widespread, the task measured part of the research, writing, and

organizing skill. Afterwards a tool could produce it in minutes: a convincing introduction,
subheadings, general examples, and a coherent conclusion. The problem was not that every
use of a tool is cheating, but that the final output was no longer sufficient evidence that the

student had read, compared, or formed a judgment.

The course team began by reworking the question from the end: what was the real learning
outcome? The aim was not to produce an attractive essay, but for the student to distinguish

a general claim from evidence, to connect the effect of the technology to a local context,
and to justify a professional recommendation. On that basis the task was redesigned into
four short stages instead of a single submission. In the first stage the student writes a 150-
word initial prediction before using any tool. In the second they choose two approved

sources and complete a short source card: what is the idea? What are the limits of the
evidence? And what does the source not establish?

In the third stage, AI use is permitted with conditions: generating revision questions,
proposing a structure, or testing the strength of the argument. It may not write the final text
in full. The student attaches a six-line disclosure naming the tool, the purpose, and what they

275