Welcome to Module 6: Ask Who Could Be Affected
Work through the explanation, apply it to the example, and complete the quick check before continuing.
A useful result can still create harm
A fictional training centre wants to use AI to sort course applications. The system would read forms, rank applicants, send rejection messages, and create accessibility notes for accepted learners.
Before asking whether the output is accurate, ask:
- Whose information is used?
- Who may gain or lose an opportunity?
- Could any group be treated unfairly?
- Can every learner use the process?
- What can the system read, change, or send?
- Who has authority to decide and stop it?
Responsible use is not a final warning added after a system is built. It begins with the task, people, information, permissions, and possible consequences.
What you will learn
You will learn to:
- Minimise personal and confidential information.
- Treat external text, files, links, and tool results as untrusted content.
- Recognise prompt injection and added risks when systems can act.
- Consider bias, fairness, accessibility, and affected people.
- Identify copyright, licence, attribution, disclosure, and ownership questions that need evidence or specialist review.
- Name a responsible owner, meaningful human decision, escalation route, and stop condition.
This module gives practical introductory controls. It is not legal, security, privacy, or professional advice. Organisational policy, applicable law, and qualified specialists take priority.
Main rule
If you cannot name the affected people, permitted data, responsible owner, and stop condition, the work is not ready to proceed.
We will begin with the information the system receives.