Page 388 - CyberSecurityEssentialsEnglish

Preventing Skill Loss and Over-Reliance
A tool may raise a beginner’s performance quickly, but it may also stop them building the
mental model needed to review the results. And it may make an expert faster while reducing

their attention if the automated recommendation always becomes the starting point.
Periods and tasks that preserve human practice should therefore be designed: writing a draft
without a tool, solving a sample by hand, explaining why a result was accepted, or conducting

a blind review that does not see the system’s recommendation first.

✓ Practices that preserve skill
• Ask for an independent judgment before showing the system’s recommendation in
sensitive tasks.
• Set aside periodic samples completed without a tool, and train teams on real failure
cases.

• Measure quality and rework, not only the number of outputs or the speed.

◇ Case study: putting the tool in the right place
This composite teaching case is provided for explanation and does not describe a specific
event.

Background: a customer service department wanted to use a text generator to increase
reply speed.
Problem: drafting time fell, but some staff began relying on the suggested reply and
reading it quickly.
Decision: the team redesigned the process so that the system presents points grounded
in an approved policy and asks the employee to choose the resolution and justify it, with
weekly sample review.
Result: the time saving continued and reopened cases fell.

Lesson: the value does not come from replacing judgment, but from placing the tool in
the right position.

Redesigning Jobs Rather Than Deleting or Freezing Them

When AI is introduced, effort may shift from producing a draft toward defining the problem,
preparing the inputs, reviewing, and communicating with the beneficiary. This shift may
raise the value of a role, but it may also increase cognitive load if review is added to the old
workload without adjusting the targets and the time. Actual load must therefore be

measured, rather than assuming that every minute the tool saved has become free capacity.

It is important to analyze how the gains are distributed. Management may benefit from
increased output while staff carry the burden of correcting errors or monitoring exceptions.

Routine tasks that were once an entry point for a beginner’s learning may be moved away,
making it harder to build expertise. Fair design reviews roles, career paths, training,
compensation, and performance indicators, not the tool alone.

387