Skip to the lesson
AI LiterateEthical Dilemmas in AI lesson

This lesson is being checked by experts. How we check lessons

Step 2 of 6Explainer7 min

Four questions for hard calls

Three ideas help you judge an AI tool.

: Treating people in an equal and reasonable way.

: Someone is answerable when things go wrong.

: People can find out how a decision was made.

Four questions help you use them.

  1. Who is affected?
  2. What could go wrong?
  3. Who is responsible?
  4. What would make it fairer?

Tip

Start with question 1. You can’t judge a tool until you know who it touches.

Try them on a made up case. A grocery chain gets 500 job applications. It uses a tool to rank the resumes. The tool learned from past hires, and most of them came from one group. So it ranks people like them higher.

Who is affected? Everyone who applies, and the store.

What could go wrong? Strong applicants from other groups get ranked low and are never seen.

Who is responsible? The chain that chose the tool, and the company that built it.

What would make it fairer? A person reviews the low ranked resumes, and the chain tells applicants it uses a tool.

Think about it

Which of the three ideas does this tool put at risk? Pick one and say why.