Deliverable Overview
You'll write a position map of 700 to 1,000 words on one AI controversy that is live when you write. Your reader is a friend or colleague who has asked what to make of a story in the news and expects a straight answer. If the map favors one side by how it describes the other, they'll notice. If it treats every claim as equally supported, it won't help them.
Nobody grades the map. You check it yourself against the success criteria listed here, and you keep it.
Detailed Requirements
The map has six parts, in this order.
The question
State the controversy in one sentence, as a question with more than one defensible answer. Keep it narrow enough that evidence could bear on it. "Is AI good or bad for workers?" is too broad to map. "Has AI reduced the hiring of young workers in the United States?" can be mapped.
The positions
Give at least two positions. Write each in three to five sentences that its advocates would accept. Attribute each to a named person or document, and give the person's role and affiliation.
Then state the advocate's stake: what they or their employer gain or lose depending on which answer people believe. A stake can be money, reputation, or an institution's mission. Having a stake doesn't make a position wrong. Your reader is entitled to know about it.
The evidence
For each position, give its best evidence, usually one or two items, and label each item by kind.
| Kind | What it is | What it can support |
|---|
| Laboratory test | A study in which researchers build the conditions themselves, as opposed to observing what happens in ordinary use | What a system does under those conditions |
| Survey | A study that asks a sample of people the same questions and reports the share giving each answer | What people say they do or think |
| Field data | Records of what happened in ordinary use, such as payroll, sales, or usage records | A pattern in the real world, often without its cause |
| Court ruling | A court's decision on one dispute | What the law requires in that case and jurisdiction |
| Forecast | A statement about what will or may happen, which can't be checked until the time arrives | What its author expects, and on what assumptions |
If an item fits none of these, describe it in your own words.
Where the evidence stands
Separate three things: findings that both sides accept, interpretations of those findings that the sides dispute, and forecasts. Then say where the evidence is stronger, weaker, or absent. If it's lopsided, say so.
Your view
State your current answer to the question, your confidence in plain words, and one or two specific things that would change your mind.
The sources
List every source with its author, date, title, and a named link. A named link shows the name of the site or document as the link text in place of a web address. In-text citations are welcome, in the form (Author Year).
Success Criteria
Your map is complete when:
- The controversy is stated as a question
- At least two positions are stated as their advocates would accept
- Each position names an advocate, with role and stake
- Each piece of evidence is labeled by kind
- You said where the evidence is stronger, weaker, or absent, without forcing a balance
- You stated your view, your confidence, and what would change it
- Every source is listed with a named link
Step-by-Step Guidance
The steps use one running example, a dispute over entry-level jobs, described as it stood in October 2026. The example shows the method. Choose a different controversy for your own map.
Step 1: Choose a controversy that's live now and state it as a question with more than one defensible answer
Start from a story you've seen in the past few weeks. Find the claim under dispute and turn it into a question. The example question: has AI reduced the hiring of young workers in the United States?
Step 2: Find a primary statement of each position by someone who holds it
A primary statement is a paper, essay, ruling, official statement, or transcript by the person or body that holds the position. A news article about a position can lead you to one. The example uses the published summary of a Stanford University study and a critique of that study published by the Economic Innovation Group, a policy organization.
Step 3: Write each position in three to five sentences, with the advocate's role and stake
Position A. Three researchers at Stanford University's Digital Economy Lab report a warning sign in payroll records. Employment of workers aged 22 to 25 in AI-exposed occupations stands 19 percent below where it would be had it kept pace with that of less-exposed peers, and experienced workers show no comparable gap. The authors call these "early, descriptive indicators" and offer them "rather than causal estimates" (Brynjolfsson, Chandar, and Chen 2026). Stake: professional reputation, since the finding carries their names.
Position B. Two economists at Google argue that interest rates explain the pattern. They attribute it to "the sharpest monetary policy tightening cycle in four decades," which began in March 2022. Job postings in the most AI-exposed occupations peaked that spring, months before the first widely used AI chatbot was released in November 2022 (Iscenko and Curto Millet 2026). Stake: their employer develops and sells AI systems and benefits if AI isn't blamed for job losses.
Position A includes its own caveat because its authors do. Leaving the caveat out would turn their claim into one they haven't made.
Step 4: List each position's best evidence and label its kind
Position A rests on field data: payroll records covering millions of United States workers through June 2026. Position B rests on field data too, in the form of job postings, and on an interpretation of their timing. A third source helps with both. The Budget Lab at Yale, a policy research center, analyzed a government survey of households and reported "no statistically or economically significant effects as of yet." It describes that data as "somewhat underpowered" for small groups such as recent graduates (Gimbel, Kendall, and Nunn 2026).
Step 5: Write where the evidence is stronger, weaker, or absent
In the example, the evidence is stronger on the pattern. Both sides accept that the gap appears in the payroll records, so presenting its existence as an open question would force a balance the evidence doesn't support. The evidence is weaker on the cause, which is the point in dispute. A causal estimate, meaning a result that measures how much one thing changed another with other explanations ruled out, is absent. Any statement about entry-level jobs in the coming years is a forecast and belongs under that label.
Step 6: State your view, your confidence, and what would change it; add sources with named links
A workable form is: "My current view is… My confidence is low, moderate, or high, because… I'd change my view if…" Make the last part specific. In the example, a study comparing similar firms that did and didn't adopt AI tools would bear on the cause. "More research" is too vague to count.
Resources
- Primary sources are the base of the map: reports from governments and international bodies, court rulings, research papers, and essays or interview transcripts by the people who hold each position.
- Your own notes or practice writing on any AI debate can supply a topic or a first draft of a position.
- An AI assistant can help you find primary statements or test whether your statement of a position is fair. Open every source yourself. An assistant's summary isn't a source, and assistants sometimes misdescribe or invent references. Don't paste in anything confidential.
Common Pitfalls
- Stating one side in its own words and the other in its critics'. Take each position from a person who holds it.
- Citing a forecast as a finding. A prediction about the future is evidence of what its author expects.
- Leaving out who benefits. State the stake for every advocate, including the ones you agree with.
- Splitting the difference when the evidence is lopsided. Where one claim is far better supported, say so.
- Hiding your own view. Your reader asked for a straight answer. Give one, with your confidence.
References
- Brynjolfsson, Erik, Bharat Chandar, and Ruyu Chen. 2026. "Canaries in the Coal Mine? Six Facts About the Recent Employment Effects of Artificial Intelligence." Stanford Digital Economy Lab. First released August 2025; revised August 12, 2026.
- Gimbel, Martha, Joshua Kendall, and Ryan Nunn. 2026. "What We Do and Don't Know About How AI Is Affecting the Labor Market." The Budget Lab at Yale, May 7, 2026.
- Iscenko, Zanna, and Fabien Curto Millet. 2026. "Looking for the Ladder: Is AI Impacting Entry-Level Jobs?" Economic Innovation Group, January 2026.