Lesson 2 of 12

Structured learning draft

Harms with Safety

In Responsible AI Development, the way a learner handles harms shapes how Safety is used and evaluated. Harm analysis considers severity, scale, duration and distribution. This intermediate lesson focuses on a decision or output that another person can inspect.

Learning objectives

Explain harms in the context of Responsible AI Development.
Apply Safety to a bounded practical task.
Evaluate the result using explicit quality criteria.

Instructional figureDecision matrix

Harms: from context to evidence

Harms connects affected people and context to a documented mitigation in Responsible AI Development.

Affected people and context

Define the purpose, intended user and Safety constraints.

frames

Harms

Write concrete scenarios naming affected people.

produces evidence for

Documented mitigation

Compare the observed result with a normal case, boundary case and stated limitation.

Takeaway: Harms is credible only when the result can be traced back to its purpose, inputs and constraints. Harms is the decision layer between the starting context and evidence that the result is fit for purpose.

Harm analysis considers severity, scale, duration and distribution. For Safety, distinguish performing an operation from demonstrating that it suits the stated purpose. Write concrete scenarios naming affected people. Record assumptions that could change the conclusion.

Apply harms deliberately

State the Responsible AI Development task and the decision it supports.
Prepare a small Safety case with a known input and difficult boundary.
Write concrete scenarios naming affected people.
Compare the observed result with the expected behaviour and explain differences.
Save the evidence, limitation and next action in a review record.

harms evidence record
Review point	Evidence
Purpose	The specific Safety outcome and intended user
Method	The harms decision, input and version or context
Result	Observed output plus a checked boundary case
Limitation	What the result does not establish and the next safe action

Common mistakes

Using Safety before defining what harms must achieve.
Checking only the easiest Responsible AI Development example.
Reporting a result without its input, assumptions or limitation.

Practice activity

Apply the lesson

For Responsible AI Development, complete a bounded Safety task demonstrating harms. Keep the original input, numbered method, normal test, boundary test, observed results and a 100-word self-review naming one limitation and next improvement.

Check your understanding

In Responsible AI Development, which evidence best supports a harms result produced with Safety?

A polished Safety output without its input The input, expected behaviour, normal and boundary results, and a limitation A screenshot showing the tool opened A claim that the method always works

Lesson summary

For Responsible AI Development, harms means: Harm analysis considers severity, scale, duration and distribution.
A credible Safety result includes a checked boundary, not only a successful example.
The next lesson builds on this harms evidence record.

Sources and further reading

Recommendation on the Ethics of AIUNESCO - accessed 2026-08-21
AI PrinciplesOECD - accessed 2026-08-21

Personal study note

Your notes

Saved on this device

Notes for this lesson

0 / 12,000No saved note yet.