Example · ELA · Grade 7 · United States

Fact-Checking AI: Birmingham History Detectives

Keisha, Birmingham, Alabama

Built by ALF Studio from exactly what each teacher typed, then changed the way that teacher teaches.

What she typed · word for word

“Seventh graders ask three AI chatbots the same questions about Birmingham history, then check each claim against primary sources at the Birmingham Public Library archives and the Civil Rights Institute. They write a guide to catching AI mistakes and teach it to fourth graders.”

On the form: United States (Common Core and NGSS) · Grade 7 · ELA · 4 weeks · 5 lessons a week, 50 minutes · Birmingham, Alabama

Why it is here

Students catch chatbots getting their own city's history wrong, then become the teachers.

What a school would ask first

The district's AI policy and age limits (the teacher runs prompts on a projected account if needed); civil rights history includes violence, so frame it with the Institute's education staff.

Example teacher

Keisha

Birmingham, Alabama

Teaches: Grade 7 ELA, United States (Common Core and NGSS).
Her lessons: 5 a week, 50 minutes.

Standards on the form

W.7.9: Draw evidence from literary or informational texts to support analysis, reflection, and research.Suggested by ALF, kept
W.7.1: Write arguments to support claims with clear reasons and relevant evidence.Suggested by ALF, kept
RI.7.8: Trace and evaluate the argument and specific claims in a text, assessing whether the reasoning is sound and the evidence is relevant and sufficient to support the claims.Chosen with Choose your own

The unit as she left it

Drafted Oct 8 in 22 seconds · 5 changes since

ELA · Grade 7 · 4 weeks

Fact-Checking AI: Birmingham History Detectives

Driving question

How can we teach fourth graders to spot AI errors about Birmingham history using local primary sources?

Standards

W.7.9

Draw evidence from literary or informational texts to support analysis, reflection, and research.

2 parts

aApply grade 7 Reading standards to literature (e.g., "Compare and contrast a fictional portrayal of a time, place, or character and a historical account of the same period as a means of understanding how authors of fiction use or alter history").
bApply grade 7 Reading standards to literary nonfiction (e.g. "Trace and evaluate the argument and specific claims in a text, assessing whether the reasoning is sound and the evidence is relevant and sufficient to support the claims").

TAUGHT IN Week 1, activity 3; Week 2, activity 1; Week 2, activity 2; Week 2, activity 3; Week 3, activity 2; Week 4, activity 4ASSESSED BY Archival Evidence

W.7.1

Write arguments to support claims with clear reasons and relevant evidence.

5 parts

aIntroduce claim(s), acknowledge alternate or opposing claims, and organize the reasons and evidence logically.
bSupport claim(s) with logical reasoning and relevant evidence, using accurate, credible sources and demonstrating an understanding of the topic or text.
cUse words, phrases, and clauses to create cohesion and clarify the relationships among claim(s), reasons, and evidence.
dEstablish and maintain a formal style.
eProvide a concluding statement or section that follows from and supports the argument presented.

TAUGHT IN Week 3, activity 1; Week 3, activity 2; Week 3, activity 3; Week 4, activity 1; Week 4, activity 2; Week 4, activity 3ASSESSED BY Argument Writing

RI.7.8

Trace and evaluate the argument and specific claims in a text, assessing whether the reasoning is sound and the evidence is relevant and sufficient to support the claims.

TAUGHT IN Week 1, activity 1; Week 1, activity 2; Week 1, activity 3; Week 2, activity 2; Week 3, activity 3ASSESSED BY Evaluating Claims

Overview

Seventh graders test how well AI chatbots know Birmingham history by prompting three different tools with local civil rights questions and highlighting conflicting claims. Working in pairs, they cross-examine these claims against digitized primary sources from the Birmingham Public Library archives and Birmingham Civil Rights Institute collections. Students then write an evidence-based field guide showing how to verify digital claims and teach fourth graders strategies to spot AI errors.

Entry event

The teacher displays three short AI chatbot responses to the question, 'What happened during the 1963 Children's Crusade in Birmingham?' Side-by-side, students notice that two chatbots give conflicting dates, invent participant names, and omit key local leaders. The teacher asks the class how fourth graders doing school research would know what is true.

Need to know

Why do AI chatbots make up facts about local history?
Where can we find trusted primary sources from Birmingham archives?
How do we prove whether a claim in an AI text is supported or false?
How can we explain fact-checking steps so a fourth grader understands them?

Week by week

Your lessons: 5 a week, 50 minutes.

Week 1

Test three AI tools on Birmingham history and list their specific factual claims.

Analyze three sample AI outputs about local events and sort claims into facts, gaps, and suspected errors.
· 1 lesson
Develop five shared research questions about Birmingham civil rights events and enter them into three AI chatbots. Start with questions about the 16th Street Baptist Church and Kelly Ingram Park, since students can walk there and read the historical markers themselves.
· 2 lessons
Compile chatbot answers into a side-by-side comparison chart and isolate specific factual claims to verify.
· 2 lessons

Supports

Provide sentence frames and claim-identification highlighters for multilingual students and students with IEPs; encourage students ready for more to test how small changes in prompt wording alter the chatbot claims.

Deliverable

Side-by-side comparison chart of chatbot claims with highlighted verification targets.

Week 2

Investigate primary sources from Birmingham archives to test chatbot claims.

Examine digitized archival photographs, letters, and oral histories from local library and institute collections.
· 2 lessons
Evaluate whether archival evidence supports, refutes, or adds missing context to each chatbot claim.
· 2 lessons
Organize primary source citations and log verified facts versus confirmed AI errors in a research journal.
· 1 lesson

Supports

Offer audio recordings of oral histories and guided text excerpts for multilingual students and students with IEPs; have advanced students compare multiple primary source accounts that offer differing perspectives on the same event. Give students a claim-tracking sheet with columns for 'chatbot said', 'source says', 'match, partial, or wrong' and 'where I found it'.

Deliverable

Fact-checking research log citing specific primary source evidence for each claim.

Week 3

Draft the fact-checking guide and build step-by-step verification arguments.

Model how to write an argument proving an AI claim is false using direct primary source evidence.
· 1 lesson
Draft a fact-checker guide comparing errors from two chatbots, showing verified case studies, and outlining rules for young researchers.
· 2 lessons
Conduct peer review using a rubric checklist focused on clear claims, source evidence, and age-appropriate language.
· 1 lesson

Supports

Supply structural graphic organizers and transition word banks for multilingual students and students with IEPs; challenge advanced writers to include counter-arguments explaining why an AI might have generated that specific error.

Deliverable

Complete draft of the student-authored AI fact-checking guide.

Week 4

Revise the guide into a teaching tool and present it to fourth-grade students.

Revise and finalize the guide based on peer feedback, polishing arguments and visual layout.
· 2 lessons
Rehearse mini-lessons in pairs, practicing how to guide fourth graders through testing an AI claim.
· 1 lesson
Deliver the fact-checking workshop to fourth graders and guide them through an interactive claim check.
· 1 lesson
Complete a written reflection assessing personal growth in evaluating digital and archival texts.
· 1 lesson

Supports

Provide cue cards and paired presentation roles for multilingual students and students with IEPs; invite advanced students to field spontaneous research questions from the fourth-grade audience.

Deliverable

Final polished fact-checking guide and completed self-reflection form.

Critique and revision

The end of week 3, period 3.

Pairs swap guide drafts and complete a two-stars-and-a-wish rubric review, focusing on whether claims are backed by specific archival evidence and whether the instructions are easy for fourth graders to follow.

Student reflection

Students complete a written exit reflection analyzing how checking primary sources changed their trust in online tools and what skills they gained as researchers.

Final product

What they make

An illustrated fact-checking guide and a mini-lesson teaching younger students how to verify AI claims using primary sources.

Options

Pairs choose whether to make their guide as a printed booklet, a laminated cheat sheet, or a slide deck, and they select which verified Birmingham history case study to highlight. Let teams choose their teaching format for the fourth graders: a comic strip, a short skit, or a game of 'spot the fake fact'.

Who sees it

Fourth-grade elementary students during a workshop visit in the final week of the unit.

Teams

Teams of two, formed by student choice with teacher approval to ensure balanced strengths.

Rubric3 rows · 4 levels

Criterion	beginning	developing	proficient	advanced
Evaluating Claims RI.7.8 · individual	Identifies chatbot claims but does not assess whether the reasoning or supporting evidence is valid.	Identifies claims and points out errors, but provides incomplete explanations of why the reasoning fails.	Accurately traces chatbot claims and evaluates whether their historical statements are sound and supported.	Critically assesses subtle inaccuracies or bias in chatbot reasoning and clearly explains the limits of its evidence.
Archival Evidence W.7.9 · individual	Mentions primary source materials without quoting or connecting specific details to the claims.	Includes some details from primary sources, but the evidence is loosely connected to the investigated claim.	Draws accurate, relevant evidence from Birmingham archive materials to support factual analysis and claims.	Skillfully weaves specific, compelling textual evidence from multiple primary sources into research analysis.
Argument Writing W.7.1 · team	States a claim about an AI error but lacks logical reasons, organized evidence, or formal tone.	Constructs an argument with a clear claim, but reasons are unevenly supported and transitions are weak.	Writes clear arguments with logical reasons, accurate evidence from sources, and an appropriate formal style.	Produces a compelling, cohesive argument that thoroughly refutes false claims and anticipates counter-arguments.

Evaluating Claims

RI.7.8 · individual

beginning

Identifies chatbot claims but does not assess whether the reasoning or supporting evidence is valid.

developing

Identifies claims and points out errors, but provides incomplete explanations of why the reasoning fails.

proficient

Accurately traces chatbot claims and evaluates whether their historical statements are sound and supported.

advanced

Critically assesses subtle inaccuracies or bias in chatbot reasoning and clearly explains the limits of its evidence.

Archival Evidence

W.7.9 · individual

beginning

Mentions primary source materials without quoting or connecting specific details to the claims.

developing

Includes some details from primary sources, but the evidence is loosely connected to the investigated claim.

proficient

Draws accurate, relevant evidence from Birmingham archive materials to support factual analysis and claims.

advanced

Skillfully weaves specific, compelling textual evidence from multiple primary sources into research analysis.

Argument Writing

W.7.1 · team

beginning

States a claim about an AI error but lacks logical reasons, organized evidence, or formal tone.

developing

Constructs an argument with a clear claim, but reasons are unevenly supported and transitions are weak.

proficient

Writes clear arguments with logical reasons, accurate evidence from sources, and an appropriate formal style.

advanced

Produces a compelling, cohesive argument that thoroughly refutes false claims and anticipates counter-arguments.

Materials

Digitized primary source collections from local public library archives and civil rights museum records
check before use
Classroom internet access to three approved school-safe AI chatbot platforms
check before use
Sample printouts of chatbot outputs containing known historical inaccuracies
Highlighters, graphic organizers for claim tracking, and guide template paper

Week 4, ready to teach

5 lessons · Built Oct 8 in 88 seconds

The week card on her unit reads: "Started from your changes to weeks 1, 2 and 3: Made it local, Added a scaffold, Trimmed it." Each lesson is a page for her, the task sheet is for the class, and ALF wrote the attachments the lessons hand out.

Week 4 · Lesson 1 of 5

Simplifying Sentences for Young Detectives

Minutes50

Prep5 minutes

Parts4

BuiltOct 8

Checkpoint

Each student writes revised text arguments rewritten in age-appropriate sentence structures on their draft fact-checking guide, confirmed by the teacher checking their rewritten claim comparisons.

Objective

Students will revise informational text arguments to support claims with clear reasons, adapting sentence structures and replacing passive constructions with active voice for a younger audience.

Need to know

How do we rewrite complex archival arguments so fourth graders can understand our evidence?

Sequence, 50 minutes39 of 50 minutes planned. The rest is yours.

Hook · 5 min

Display two versions of an argument refuting a chatbot error about the 1963 Children's Crusade. Ask students to contrast which version an elementary reader can follow quickly and point out the specific differences.

Mini-lesson · 9 min

Explain that writers adapt informational syntax for younger audiences by replacing passive constructions and multi-clause sentences with direct subject-verb-object structures and concrete verbs. Show how dense passive clauses obscure who acted and confuse readers. Worked example, with example figures: The passive claim 'The demonstration was organized by students who were told by leaders that leaving school would show courage' becomes 'Birmingham students organized the march. They left school to demand civil rights.' Define the target sentence word count as an average limit to keep focus sharp.

Work time · 20 min

Both partners analyze their draft fact checking guide to identify multi-clause sentences and passive verbs in their case studies. Each student rewrites at least three dense claim comparisons into direct subject-verb-object structures that meet the target sentence word count, creating the simplified explanatory text and revised claim arguments directly in their guide draft. The teacher circulates to review sentence structures and ensure active voice preserves historical accuracy.

Check and close · 5 min

Collect each student's revision index card listing one original passive sentence beside their final simplified version. For students whose revised sentences remain over-complicated or passive, pull them during the opening of lesson 2 to provide sentence-combining frames before layout work begins. Next lesson, pairs format these polished arguments into the printed booklet layout.

Materials

havethe draft fact checking guide

haveIndex cards for sentence check

makeSyntax Revision Guide showing passive vs active sentence pairs

Watch for

Students may cut critical historical evidence when trying to shorten sentences. Remind them to split one long thought into two short sentences rather than deleting facts.

Differentiation

Scaffold.Writing load barrier: provide structural graphic organizers and sentence frames with labeled subject-verb-object slots for students with IEPs and students learning English.

Extension.Challenge advanced writers to include counter-arguments explaining why an AI might have generated that specific error using direct, simplified syntax.

Next class

Teams place their simplified arguments into the four-page booklet template with archival captions.

Her three edits

Each labeled by ALF after the save

What Keisha changed in the draft, in her own words, with the label ALF gave each change and the reason it suggested.

Made it localWeek 1 · Activity 2Probably: local context
Develop five shared research questions about Birmingham civil rights events and enter them into three AI chatbots. Start with questions about the 16th Street Baptist Church and Kelly Ingram Park, since students can walk there and read the historical markers themselves.
Added a scaffoldWeek 2 · SupportsMaybe: support
Offer audio recordings of oral histories and guided text excerpts for multilingual students and students with IEPs; have advanced students compare multiple primary source accounts that offer differing perspectives on the same event. Give students a claim-tracking sheet with columns for 'chatbot said', 'source says', 'match, partial, or wrong' and 'where I found it'.
Changed what students makeFinal product
Pairs choose whether to make their guide as a printed booklet, a laminated cheat sheet, or a slide deck, and they select which verified Birmingham history case study to highlight. Let teams choose their teaching format for the fourth graders: a comic strip, a short skit, or a game of 'spot the fake fact'.
The catalogue expected Gave students a choice; ALF read it as Changed what students make and that label stands.

Then she pressed Adapt this week

Week 3 · Running behind

Her note, with Running behind picked

“Archive visit ate two days. Cut the third chatbot and compare only two, so drafting the guide starts Wednesday.”

ALF suggested. Week 3 shortened to three days.

Kept · ChangedTrimmed it
Conduct peer review using a rubric checklist focused on clear claims, source evidence, and age-appropriate language. · 2 lessons
Conduct peer review using a rubric checklist focused on clear claims, source evidence, and age-appropriate language. · 1 lesson
One day for peer review keeps the week within three days after the archive visit.
Kept · ChangedTrimmed it
Draft a four-section fact-checker guide explaining AI errors, showing verified case studies, and outlining rules for young researchers.
Draft a fact-checker guide comparing errors from two chatbots, showing verified case studies, and outlining rules for young researchers.
Focusing on two chatbots fits drafting into the remaining days.

She kept 2 of 2.

Then ALF built week 4. The week card reads “Started from your changes to weeks 1, 2 and 3: Made it local, Added a scaffold, Trimmed it.” Read the lessons.

Build one from your own idea

Type what you want to teach the way Keisha did. ALF drafts the unit, and you change whatever does not fit your class.

Start a unit All examples