Stakeholders and Ethical Dilemmas

Think beyond the user

A hiring screener affects applicants who never chose to use it. A support bot affects agents whose jobs change. A recommender affects communities. List every affected group, ask what each gains and risks, and notice whose voice is missing. Real dilemmas arise when goals conflict: accuracy vs explainability, personalisation vs privacy, efficiency vs jobs, automation vs human judgement. There is rarely a perfect answer; the aim is a reasoned, documented, reviewable choice made by accountable people.

A stakeholder table

Fill it for every high-impact use case. The "Voice" column forces you to say how that group is heard.

Group              Benefit                 Risk                       Voice
Applicants         faster response         unfair screening           appeal route + human review
Recruiters         less manual sorting     over-trusting scores       training, override button
Hiring managers    better shortlists       missing good candidates    quarterly sample audit
Society/regulators consistent process      systematic bias            published summary, audits

Quick check: Who might be affected by an AI hiring screener besides its users?

• Job applicants who are screened by it
• Nobody
• Only the developers
• Only the cloud provider