Lecture 3-5: Practice Lab: Who Is Missing?
Contents 31
- Lecture 1-1: Why AI Ethics, and Why Now?
- Lecture 1-2: The Image of God and the Dignity of Persons
- Lecture 1-3: Three Traditions, One Set of Questions
- Lecture 1-4: Wisdom for Hard Cases: Five Questions
- Lecture 1-5: Practice Lab: The Five Questions
- Lecture 2-1: Who Is Responsible? The Gap
- Lecture 2-2: How Scripture Speaks of Responsibility
- Lecture 2-3: Decisions a Person Must Keep
- Lecture 2-4: Admitting Error and Repairing
- Lecture 2-5: Practice Lab: A Map of Responsibility
- Lecture 3-1: Where Bias Comes From
- Lecture 3-2: What Is Fair? Scripture and Several Measures
- Lecture 3-3: Structures and Persons
- Lecture 3-4: Who Is Missing? Listening in Church and Work
- Lecture 3-5: Practice Lab: Who Is Missing?
- Lecture 4-1: The God Who Sees and the Watchers
- Lecture 4-2: Privacy, Consent and Entrusted Information
- Lecture 4-3: Safety and Liberty: Uses of Watching
- Lecture 4-4: Being Watched and Watching with Love
- Lecture 4-5: Practice Lab: What Is Seen, What Is Hidden
- Lecture 5-1: A Theology of Work
- Lecture 5-2: Automation: History and What We Do Not Know
- Lecture 5-3: Dignity, Sharing and the Neighbour Who Loses Work
- Lecture 5-4: Rest, Worth and Not Being Useful
- Lecture 5-5: Practice Lab: My Work, My Neighbour's Work
This is the third Practice Lab. Every week of this course ends in one, and each asks you to judge a case, not only to agree with a teaching.
Today's task is small and exact. Take one invented selection process and find who is missing from it. Lecture 3-1 gave you three places where unfairness enters. Lecture 3-2 gave you four measures of fairness. Lecture 3-3 gave you the structures-and-persons test. Lecture 3-4 gave you seven habits of listening. The lab puts them on one sheet.
You are not asked to find the right answer, because in a case like this there is rarely only one. You are asked to see more than you saw at first glance, and to hold your judgement with humility.
What these 19 minutes give you
- Complete the "Who Is Missing?" form on an invented selection process.
- Name one example of each of the three sources of bias in a case.
- Say which measure of fairness a process seems to use and whom it leaves exposed.
- Choose one person or group you would ask, and say how you would ask them without harming them.
- Explain the four privacy promises of the lab and keep them in a group.
for I was hungry, and you gave me food to eat. I was thirsty, and you gave me drink. I was a stranger, and you took me in." — Matthew 25:35
World English Bible
I was naked, and you clothed me. I was sick, and you visited me. I was in prison, and you came to me.’" — Matthew 25:36
WEB
Then the righteous will answer him, saying, ‘Lord, when did we see you hungry, and feed you; or thirsty, and give you a drink?" — Matthew 25:37
WEB
When did we see you as a stranger, and take you in; or naked, and clothe you?" — Matthew 25:38
WEB
When did we see you sick, or in prison, and come to you?’" — Matthew 25:39
WEB
The King will answer them, ‘Most certainly I tell you, because you did it to one of the least of these my brothers, you did it to me.’" — Matthew 25:40
WEB
The people in the story are surprised. "When did we see you?" They had not known whom they were serving. The passage does not say that seeing is easy. It says that those who did see the hungry and the stranger, and acted, were blessed, and it leaves us asking whom we have not seen. That is the question of this lab. Keep in mind that this is a parable about the judgement, not a rule for a selection process. We will use it only as a reminder that the least are not always in view.
1. What this lab is for
A case is a safe place to practise. In a real committee, the pressure of time, friendship and the wish to be seen as fair can make it hard to see who is missing. In a printed case no one is embarrassed, and you can try a question and then discard it.
The lab does three things. It slows you down, so you do not stop at "the list looks reasonable." It gives you the questions in a fixed order, so that they become a habit. And it lets you notice where you tend to stop looking.
2. The three ways in
Everyone can do this lab. Choose one of the three ways below, or use two.
Three ways into the lab
1. In your own place. If you serve on a committee, run a programme, select volunteers, give out help, or have to rank or choose people in any way, think of one kind of choice you or your group make. Do not write any name, any group of real people, or any confidential detail on the form. Write only the shape of the process ("a sign-up list for a visiting rota", "a way of choosing who receives a bursary"). Then fill in the form from that shape.
2. The observer's way. Use the invented case printed below, and fill in the form from it. Nobody is described who exists. This is the way for anyone who does not choose people in their daily life, and it is a good way for everyone.
3. The partner's way. Sit with a trusted partner. Take turns. One reads the case aloud slowly. The other says in their own words who they think is missing, and why. Then swap roles with the questions on the form. No personal story is required. Each of you may pass on any question.
3. The invented case: the Millbrook Apprenticeship
(An invented example) A community trust in a town offers twelve places each year on a paid six-month apprenticeship in bookkeeping. About a hundred and fifty people apply. The trust has two staff members and a small volunteer board.
To save time, the trust uses an online application form with an automated first screening step. The form asks for a typed letter of five hundred words, a scan of a school or college certificate, and a reference from a recent employer. The screening step gives each application a score for "match with successful past apprentices," and the staff read only the top forty.
The screen uses three things: how close the letter's wording is to the letters of past apprentices; whether the certificate comes from one of the two nearby schools; and whether the applicant's work history has no gaps longer than three months. The trust chose these because past apprentices who did well had these features.
Last year, a volunteer on the board looked at the twelve who were chosen. Eleven were under thirty. Nine came from the two nearby schools. None had a gap in their work history. None had applied in a language other than the main one.
The volunteer asked a few people in the town why they had not applied. One woman who cares for her mother at home said she had a gap of two years and thought it was not worth trying. A man who had learned bookkeeping informally in a family shop said he had no certificate and no employer to give a reference. A young woman from a village said the website had failed twice on her phone and she had run out of data. A fourth person said she had written the letter in her second language with help from a friend and had been told it "did not read naturally."
The trust's staff are kind and tired. They say they did not exclude anyone, and the form is open to all. The board is divided: some want to keep the screen, since it saves time, and some want to drop it.
4. How to do it
Five steps
Step 1. Read the case once for the story, once for the people. On the second reading, underline every person or kind of person who is mentioned only as a reason, or not mentioned at all.
Step 2. Find the three sources. In the form, write one example from the case for each of: the recorded past, the design choices, the use of the result.
Step 3. Name the measure. Which of the four measures of fairness does the process seem to use? Which does it leave exposed?
Step 4. Ask who is missing. List at least three kinds of people the process does not see. Then ask: is this a pattern with an author, or a structure? Which persons could act, and what small step could each take?
Step 5. Choose one person or group to ask, and how. Say how you would go to them, and how you would keep them from harm. Finally, write one thing you do not know.
You have about thirty minutes, alone and in silence. There is no right number of answers.
5. Four promises of privacy
Say these out loud at the start, whether you are alone or in a group.
- Nobody reads what you write. The form is yours.
- Nothing is marked or graded. There is no pass or fail.
- You may pass on any part. Passing is a good answer.
- The leader fills in a sheet too. No one sits outside the exercise watching.
Never write the name of a real person, a real congregation, a real employer or a real selection on the form. If the exercise brings to mind a real harm (abuse, violence or danger), stop writing. Speak to a trusted pastor, priest or safeguarding person this week. If there is danger now, call your local emergency number.
6. What usually goes wrong
"I can't tell whether the process is unfair." That is an honest result. A process can lean and still be defended. The lab asks you to see who might be hurt, and to name what you do not know. It does not ask you to deliver a verdict.
"I keep wanting to blame the staff." Notice the wish and put it down. The case is built so that no one is cruel. If you find a person to blame, ask what the arrangement did that made it easy for them.
"I solved it by dropping the screen." Dropping the screen is one option. It does not remove the need for a measure of fairness, and the staff would have to read a hundred and fifty letters. Ask what you would lose and who might be hurt by that.
"Everyone I listed is a group I already know." Good. Now ask who is not on your list, as the did in lecture 3-4.
7. Holding the judgement humbly
At the end of the form there is a line for what you do not know. Write it honestly. In the Millbrook case you do not know whether the chosen twelve were the best, whether the people who did not apply would have done well, or whether the automated step or the humans did more harm. That is not a failure of the lab. It is the first fact about the case.
This is where the third week ends, and its principle stands finished. Unfairness in automated systems is old prejudice made to look like arithmetic. To see it, we look for who is missing, and we go to ask them.
8. Leader's note, for running this in a group
Privacy comes first
State the four promises first, and out loud. Before anyone picks up a pen, say: nobody will be asked to show the sheet, nobody will read it, it is not marked, anyone may pass, and I will fill in a sheet too. Say it even if you said it last week.
Fill in your own sheet with them. A leader who watches while others write turns the room into an examination.
Give thirty silent minutes. Do not fill the silence with commentary.
Then ask about what was seen, never about any person's own situation. Ask: "Who did you find was missing?" and "What was hard to see?" Do not ask: "Does this happen where you work?"
Allow anyone to pass, and move on without a pause that makes passing costly.
Do not let the group settle on one villain. If the room decides that the staff, the board or the screen is the culprit, ask: "What would a kind and tired person do next, and what would help them?"
Expect disagreement about the measure of fairness. This is a point where readers will not agree. Let each position be stated in its strongest form, and say that the course requires only that responsibility not be blurred and that the weak be kept in sight.
If a person's own story surfaces, do not press. Stay with them for a moment, then follow up privately after the meeting. If what you hear is beyond you, say so and help them find someone who can carry it. If there is danger now, the local emergency number comes first.
Diagram 1. The form: Who Is Missing?
In short
- The lab applies the week's tools to one invented selection process: three sources of bias, four measures of fairness, structures and persons, and listening.
- The aim is to see more than you first saw, and to hold your judgement humbly, not to deliver a verdict.
- Each of the three ways in is complete. The form never carries a real person, church, employer or confidential detail.
- Four promises govern the lab: nobody reads it, nothing is marked, anyone may pass, and the leader fills in a sheet too.
- A leader asks what was seen, never about anyone's own situation, and does not let the room settle on one villain.
- If a real harm comes to light, the lab stops and the student speaks to a trusted pastor, priest or safeguarding person.
| Front | Back |
|---|---|
| Purpose of the lab | To find who is missing from an invented selection process and to hold the judgement with humility. |
| The three ways in | In your own place; the observer's way with the printed case; the partner's way. |
| The four promises | Nobody reads it; nothing is marked; you may pass; the leader fills in a sheet too. |
| If real harm surfaces | Stop, speak to a trusted pastor, priest or safeguarding person, and call the local emergency number if danger is present. |
One sentence to remember Seeing who is missing is a practice, and it begins with a printed case, a quiet half hour and one honest line about what you do not know.
This week's action Fill in the form once more on a different invented process, and ask one trusted person to tell you who they think is missing.
Check yourself
Not "did you read it" but "can you explain it". Ticks are kept on this device only.
Reflection and discussion
- Which section of the form was hardest to complete, and what does that tell you about where you tend to stop looking?
- In the Millbrook case, which of the people described might be missing for more than one reason at once?
- The lab asks you to name something you do not know. How did you find writing it? Why might an honest sentence of this kind be a strength in a committee?
- What would you lose, and who might be hurt, if the trust simply dropped the screening step? What would you lose, and who might be hurt, if it kept the step unchanged?
- Why does the lab refuse to let anyone name a real person, a real church or a real employer on the form?