CHECK THIS AI ANSWER
A school library survey

STUDENT SOURCE PAGE

A student council is considering whether the school library should open during lunch. Someone has asked an AI tool to summarize a survey. Your job is to check each claim against the information provided before the summary goes to the council.

SOURCE CARD: LUNCH LIBRARY SURVEY
This school and its survey are fictional. All the information for this activity is below.
The school has 300 students. Sixty students answered the survey. Each person answered both questions once.
The survey did not collect grades or information about staffing costs.

1. Do you want the library to open during lunch?
Yes: 48 students
No: 12 students

2. What would you most want to do there? Choose one.
Read quietly: 30 students
Do homework: 20 students
Draw or write: 10 students

SAMPLE ANSWER TO CHECK
Written for this activity to imitate an AI summary. It is not output from a named AI tool. Some claims are supported; others need correcting or more evidence.
1. Most of the students who answered the survey want the library to open during lunch.
2. Reading quietly was the most common choice for what students would do there.
3. Forty percent of the respondents chose quiet reading.
4. Every student in the school wants the library to open during lunch.
5. Opening the library during lunch will improve students' grades.
6. The school can open the library during lunch without adding staffing costs.

STUDENT RESPONSE PAGE
Name:
Date:

For each claim, write S, C, or N. Explain your choice using a number or a detail from the source card. Do not decide from how confident the sentence sounds.
S: Supported. The source gives evidence for the claim.
C: Contradicted. The source gives evidence that the claim is wrong.
N: Not enough information. The source does not let you decide.

1. Most respondents want lunch access.
S / C / N:
Evidence:

2. Quiet reading was the most common choice.
S / C / N:
Evidence:

3. Forty percent chose quiet reading.
S / C / N:
Evidence:

4. Every student in the school wants it.
S / C / N:
Evidence:

5. Opening it will improve grades.
S / C / N:
Evidence:

6. There will be no added staffing costs.
S / C / N:
Evidence:

Rewrite the summary in 2 or 3 sentences. Include something the survey supports and something the council still needs to find out.
[Your summary]

A claim is not supported by this source. Does that automatically make it false? Explain.
[Your explanation]

TEACHER NOTES

WHO IT IS FOR
Grades 6 to 8. Useful in English, social studies, library lessons, or a mathematics discussion of percentages. Adjust the reading support for your class.

TIME AND MATERIALS
Plan for about 15 to 20 minutes; the timing has not been tested with a class. Students need pages 1 and 2 and a pencil. A calculator is optional. No internet, AI tool, paid subscription, or student account is required.

BEFORE CLASS
Print pages 1 and 2 for students. Keep pages 3 and 4 for yourself. Read the key, especially the difference between a claim the source contradicts and one it cannot answer.

WHAT STUDENTS PRACTICE
Match a claim to evidence, check what a number represents, notice when a summary goes beyond its source, and revise a summary without adding unsupported promises.

READ THE SOURCE TOGETHER
Spend about 3 minutes on the survey. Establish that 60 students answered, while 300 attend the school. Model an unrelated check: 'Ten respondents chose drawing or writing.' Ask students to point to the evidence.

CHECK THE SIX CLAIMS
Allow about 6 minutes for students to work alone or with a partner. Ask for the evidence as well as the letter. A student who chooses N should say what information is missing.

COMPARE AND REVISE
Use about 5 minutes to discuss disagreements and write the revised summary. Finish with the question at the bottom of page 2. Allow a few extra minutes if the percentage calculation needs explanation.

IF STUDENTS NEED SUPPORT
Read the source and claims aloud if helpful. Students can give their evidence orally or dictate it. For the percentage item, supply '30 out of 60 is 50%' if your focus is reading rather than computation.

IF YOU HAVE MORE TIME
Ask students what else the council should find out before deciding. Possible questions include whether other students would use the library, who could supervise, and what other lunch options students have. These are questions to investigate, not facts established by this survey.

TEACHER KEY

1. Supported (S)
Forty-eight of the 60 respondents answered yes. That is more than half. This claim refers only to the students who answered, not all 300 students.
No correction needed.

2. Supported (S)
Thirty respondents chose quiet reading, compared with 20 for homework and 10 for drawing or writing. The largest group is not necessarily a majority: 30 is exactly half of 60.
No correction needed.

3. Contradicted (C)
Thirty out of 60 is one-half, or 50%, not 40%.
Fifty percent of the respondents chose quiet reading.

4. Contradicted (C)
Twelve respondents answered no, so the word every is already contradicted. Also, the survey contains no answers from the other 240 students.
Forty-eight of the 60 respondents want the library to open during lunch.

5. Not enough information (N)
The survey did not collect grades. A preference survey does not establish what will happen to achievement. This prediction might or might not be true.
These results do not tell us whether opening the library during lunch would improve grades.

6. Not enough information (N)
No staffing or cost information is provided. Someone would need to check supervision, staff availability, and any added cost before making this claim.
The council still needs information about staffing and costs.

ONE POSSIBLE REVISED SUMMARY
Forty-eight of the 60 students who answered the survey want the library to open during lunch. Quiet reading was the most common activity choice, selected by half of the respondents. The council still needs information about staffing and costs, and this survey does not show whether extra library time would improve grades.

LAST QUESTION
No. A source can leave out information about a claim that turns out to be true. Students should explain that we need relevant evidence before treating that claim as established.

Look for a reason connected to the source, not just the expected letter. Accept different wording in the rewritten summary if it accurately states the results and identifies a genuine unanswered question.

This activity practices checking claims. It is not an AI detector and does not show how often any AI tool makes mistakes. Real responses can combine accurate and inaccurate information, and checking one source may not settle every claim. The source card and sample summary were invented for this lesson, so students can work without personal data or live AI access.

First edition. Not yet classroom tested.
Free to copy and adapt for your own teaching. Original EdTech Alchemy teaching material.
Version 1.0 | Checked September 14, 2026
edtechalchemy.com
