Professional learningfor Teachers, Librarians, Coaches
Lateral Reading Relay
Teams race to verify viral claims and an AI-written summary by leaving the page, not staring harder at it.
All activities Professional learning
Looking for the Checking
A looking-at-student-work protocol that asks one question of a class set: did students actually verify what the AI told them, and how can we tell?
Overview
Teachers now ask students to "check what the AI says," and most students write "I checked it." This protocol treats that assignment as data. A PLC reads a fictional set of eighth-grade research notes, describes what is actually on the page, places each sample on a verification evidence continuum, and plans the instruction the pattern calls for. The move for coaches: shift the conversation from "Are students cheating with AI?" to "What does our students' work show about how they verify, and what do we teach next?"
Objectives
- Participants will separate low-inference description of student work from interpretation and judgment.
- Participants will place student samples on a verification evidence continuum and describe the pattern across a class set.
- Participants will plan one instructional move, and the evidence it should produce, for each level of the pattern.
Materials
On paper
- Handout A: Student Work Set, Dust Bowl research notes (1 per person)
- Handout B: Verification Evidence Continuum (1 per person)
- Handout C: Protocol Record (1 per team, for the note-taker)
- Sticky notes in two colors (one for "I notice", one for "I wonder")
On screen
- Optional: the team's own de-identified student work, scanned into a shared folder
- Optional: a shared copy of Handout C for virtual PLCs
Before you start
- Read Handout A and the facilitator notes in Step 3 so you can steer the team back to evidence.
- Choose a timekeeper and a note-taker in advance, or assign them at the start.
- If the team will bring its own student work next time, tell them now: remove names and anything that identifies a student before it's shared.
Step by step
-
10–4 min
Hook
"I checked it."
Write on the board: "I checked it and it's right." Ask: "If a student wrote this, what do you actually know?" Take two or three answers. Say: "Today we look for the checking itself. Not the answer, not the effort. The evidence of the move."
Facilitator noteThe honest answer is: you know the student knows they were supposed to check. That's all.
-
24–14 min
Explore
Describe only
Everyone reads Handout A silently. On one color of sticky note, write only what is on the page: "Student C quotes a page number," not "Student C is careful." After six minutes, go around the table once; each person reads one note. The facilitator redirects any judgment: "What in the work makes you say that?" The note-taker records descriptions on Handout C.
Facilitator noteThis step feels slow. It is where the team stops seeing "good kids" and "lazy kids" and starts seeing moves.
-
314–24 min
Practice
Place each sample on the continuum
Using Handout B, pairs place each of the seven samples at a level for each criterion, then compare with the table. Tally how many samples reached Checked a claim or higher. Discuss the tricky ones: Student E checked a site that repeated the AI's exact wording; Student F didn't use AI at all; Student G found a mismatch and kept the AI version anyway.
Facilitator noteLook-for: the team treats E as a media literacy gap (circular sourcing) and G as a disposition gap (trusting the machine over a source). They need different instruction.
-
424–31 min
Debrief
What does the pattern say about our teaching?
Ask the coaching question in its data form: "If this were our class, what does this pattern say about what we taught, and what we assumed?" Push the team from student labels to instructional causes: "Did we ever model what a check looks like? Did the assignment ask for evidence of the check, or just the answer?"
Facilitator noteIn this set, the assignment said "verify two claims" but never showed how or asked students to write down what they compared. Most gaps trace back there.
-
531–42 min
Apply
Plan the next move for each group
In the last column of Handout C, the team plans one move for each level: for students with no check, a teacher think-aloud verifying one AI claim; for claimed checks, a required verification log (claim · where I checked · match or mismatch · what I changed); for circular sources, a mini-lesson on finding a source that didn't copy the first one; for mismatch but kept the AI, a discussion of why a documented source outranks a fluent answer. For each move, write what student work will look different next time.
Facilitator noteKeep the plan small enough to try within a week. The measure is the next class set, not a survey.
-
642–45 min
Reflect
Next data date
The team sets a date to bring the next set of student work and agrees on which criterion on Handout B they will tally. Each person writes one sentence: "Next time, I'll know my move worked if…"
Facilitator noteThis is the Measure and Adjust part of the coaching cycle. Put the date on the PLC calendar before people leave.
Paper or screen
Unplugged
The protocol is designed for paper: printed work, sticky notes, and a shared record. Looking at student work at a table, with everyone holding the same page, keeps the conversation grounded in evidence and makes it easy to point to a line.
Digital
Put de-identified scans of student work in a shared folder and use a shared copy of Handout C. In a virtual PLC, participants add "I notice" comments directly on the scanned work, then the facilitator groups them. If students did their verification in a digital log or document with version history, the team can look at when and how the check happened.
Does it need a screen? Paper is the better choice for the protocol itself. A digital verification log earns its place in the classroom afterward because it timestamps each check and preserves the claim, the source, and the revision together, which is richer evidence than a finished paragraph.
Evidence of learning
What you should be able to see or collect if it worked.
- Descriptions on Handout C stay low-inference (what is on the page) before any interpretation.
- The team distinguishes a claimed check from a documented check and names circular sourcing as its own problem.
- Planned moves are tied to the assignment design and each names the student evidence expected next time.
- A date and a criterion for the next look at student work are set before the meeting ends.
Adaptations
- Grades 3–5 teams
- Simplify the continuum to three levels: "Didn't check," "Checked one thing," "Found something to fix." Use a teacher-projected AI answer rather than student AI use.
- Higher Ed
- Use annotated bibliographies from a course that permits AI-assisted searching. Add a criterion for whether cited sources exist and say what the student claims.
- Librarians
- Co-facilitate with the PLC and lead the circular-sourcing mini-lesson that the Apply step usually calls for.
Standards connections
A PLC reads student Dust Bowl notes for evidence of real checking, then plans instruction on verifying AI claims against independent sources.
- TEKS
- inquiry and researchethics and laws (c)(9)Social studies skills (sources, frames of reference, bias, and evidence)TEKS sections: Technology Applications §126.17–§126.19, high school Technology Applications courses (19 TAC Chapter 126); ELAR §110.22–§110.24, §110.36–§110.39
Reflect
- How am I using data to inform and guide my coaching practices? (Coach ELE guiding question)
- What counts as data in our PLC, and where is student reasoning missing from what we look at?
- How would I know if students' verification improved, without relying on what they tell me?
Take it to your students
Revise one upcoming assignment so students must record each check (claim, source, match or mismatch, change). Bring the next class set to the PLC and tally the same criterion to see what changed.
Pairs well with
- Digital Inquiry Group: Civic Online ReasoningLessons for the circular-sourcing and lateral reading gaps this protocol often surfaces.
- Data Display Patterns field guideIdeas for showing the before-and-after tally to a PLC or leadership team.
