AI and Fake News: A 4-Case-Study Lesson for Grades 6-8

At lunch on a Tuesday, a teacher overheard two students huddled over a phone — one was showing the other a video clip of a well-known figure saying something that felt immediately off. “Did he really say that?” the second student asked. The first shrugged. “It’s on here, so.” The teacher had about 22 minutes before her next class and no lesson plan for that moment. If that scenario sounds familiar, this post is the one to bookmark.

An AI fake news lesson for grades 6-8 teaches students to recognize three types of AI-generated misinformation: fabricated text articles, deepfake video, and voice-cloned audio. The SIFT method — Stop, Investigate, Find better coverage, Trace claims — gives students a four-move verification protocol that works across all three formats. A four-case-study structure drawn from real 2024-2025 events gives students concrete practice rather than theory to memorize. The lesson aligns with CCSS.ELA-LITERACY.W.7.8 (gather and assess credibility of sources) and ISTE 1.3.b (evaluate accuracy, perspective, and credibility of information).

What Makes AI-Generated Fake News Different from Traditional Misinformation?

Traditional misinformation — a misleading headline, a selectively cropped photo, a stat pulled out of context — has been around long enough that most media literacy curricula have a framework for it. AI-generated misinformation is a different category, and the difference matters for how you teach it.

There are three distinct types students need to be able to name:

Synthetic text articles

A language model writes a plausible-looking article: real-sounding byline, convincing publication name, fabricated quotes attributed to real people. The article may not exist anywhere outside a social media post — but it looks like journalism. The layout, the tone, the sentence structure all signal “credible.” That signal is exactly what makes it worth studying.

Deepfake video

A person’s face and voice are swapped onto another body, or a person’s likeness is generated from scratch, appearing to say or do something they never did. The person in the video is real. The event in the video is not. For students who trust visual evidence by default, this category is the most disorienting.

Voice-cloned audio

Someone’s voice is cloned from existing recordings and used to deliver a message they never recorded. It sounds like the person. It matches their cadence, their accent, their register. In a short clip — the length of a robocall or a social media reel — most listeners cannot reliably tell the difference.

Here is the critical teaching point: telling students to trust their instincts about what “sounds off” or “looks fake” is not enough anymore. Recent reporting on deepfake detection accuracy, including from PBS AI Unlocked (April 2025), shows that even trained adults fail to consistently detect AI-generated content by feel alone. Students need a protocol — a sequence of moves they can run regardless of whether something feels fake. Instinct is a starting point, not a finish line.

Four Real 2025 Case Studies to Use in Your Classroom

Most AI misinformation lesson plans ask students to think abstractly about what fake news could look like. This one gives them four documented cases — pulled from 2024 and 2025 — that they can investigate, cross-reference, and argue about. The table below is designed to be projected or printed as a discussion starter.

Case Study	Type	Year	What Happened	Teacher Note
Biden robocall	Voice clone	Jan 2024	An AI-cloned audio clip mimicking President Biden’s voice was sent to New Hampshire voters, telling them not to vote in the primary election. Major news outlets confirmed the audio was fabricated.	Focus on the format, not the politics. Ask students: How was this created? How did journalists detect it? What would the right response have been to receiving this call?
Ireland presidential candidate deepfake	Video deepfake	2025	An AI-generated video falsely showed a presidential candidate making statements he never made. The case was documented by European fact-checkers.	The non-US context reduces political temperature in the room. Good for frame-by-frame video analysis — pause and look for tells. Source: fake-off.eu
Canada election broadcast fakes	Fabricated text article	2025	AI-generated articles mimicking CBC and CTV news formats spread on social media during Canada’s federal election, copying the visual layout, logo style, and editorial tone of legitimate broadcasters. Source: fake-off.eu	Show students the visual design — how far does the imitation go? Compare side-by-side with a real CBC article. What’s the same? What gives it away?
Voice-clone robocall surge	Voice clone	2023-2025	Voice-cloning technology was used in robocall fraud campaigns across multiple countries, with documented cases in North America, Europe, and Asia. Security researchers have flagged voice-clone fraud as one of the fastest-growing AI misuse categories between 2023 and 2025.	Connects to students’ personal lives — many have a grandparent or family member who received a suspicious call. Broadens the lesson beyond politics to personal safety.

Note: This case study set is current as of June 2026. AI-generated misinformation moves fast — plan to swap in one new real case per semester to keep the lesson current.

The ready-to-print version of this lesson — with all four case studies formatted for student analysis, a SIFT verification worksheet, and a teacher facilitation guide — is available as the AI Journalism and Fake News Lesson on TPT.

Applying SIFT to AI-Generated News — A Step-by-Step Walkthrough

Already covering SIFT for hallucinated research citations? The AI source evaluation lesson handles that scenario. Here, SIFT is applied to what students encounter on social media — a video, an article, a voice message that claims to show something real.

Here is how the four SIFT moves work when applied to the Ireland presidential deepfake case from the table above. Walk through these moves with the class before asking students to apply them independently.

Stop. Before sharing the clip — notice the impulse. Ask students to write down their gut reaction in ten seconds: “I believe this / I don’t believe this / I’m not sure.” Then ask: what made them want to react immediately? Naming the impulse is the first move. Students who can identify the urge to share, forward, or react have already started the verification process.

Investigate the source. Where did the video originate? What account posted it? Is that account verified, and for how long has it been active? For the Ireland deepfake, students can check whether the account has any history before the election period — newly created accounts that post a single viral video are a consistent pattern in coordinated misinformation.

Find better coverage. What are established Irish or UK news organizations reporting about this candidate? Is any credible outlet covering the video itself — either confirming it as a deepfake or treating it as authentic? A search across BBC News, The Irish Times, and Reuters takes under three minutes. If none of those outlets have touched the story, that absence is itself a data point.

Trace claims to origin. Can students find the original, unedited footage? Are there forensic reports? Fact-checking organizations such as fake-off.eu document specific cases with source links. Students trace the claim backward: who first posted this? What did they say it showed? Does the metadata match the claimed date and location?

The key facilitation note: the SIFT walkthrough is the lesson. Running the four moves together, as a class, with a real case on the projector, is more instructive than any quiz at the end. Students who run the moves once in a guided setting internalize the sequence.

Grade-Differentiated Verification Tasks for 6th, 7th, and 8th Grade

Grades 6, 7, and 8 don’t need the same depth of verification task. The table below gives a scaffold with suggested class times and CCSS anchor standards for each grade level.

Grade	Verification Task	Suggested Class Time	CCSS Anchor
6	Verify the source or person exists: Can you find this person or organization in three trusted places online? Students document each source with a URL and one sentence on why it’s trustworthy.	20-25 min	CCSS.ELA-LITERACY.W.6.9
7	”Does the source say what was claimed?” — trace the original quote or video back to its earliest source. Add a second question: Who benefits from sharing this, and who might be harmed?	30-35 min	CCSS.ELA-LITERACY.W.7.8
8	Full SIFT audit: complete all four moves, run a reverse image search on any still frame from the video, cross-reference findings with a known outlet. Submit a 3-sentence verification report: what the claim is, what evidence was found, and a final call — real, fake, or unresolved.	40-45 min	CCSS.ELA-LITERACY.RI.8.6

Adjust the grade-7 and grade-8 tasks for below-level readers by pairing students and providing a sentence stem for the verification report. For high-achieving classes, add a fourth column: “What question does your SIFT audit leave unanswered?”

Standards Crosswalk: Where This Lesson Fits Your Curriculum

Submitting a lesson plan for administrator review? Here is where the AI fake news unit maps across the standards frameworks most departments already use.

Standard	Name	How This Lesson Meets It
ISTE 1.2.b	Digital Citizen — engage in ethical, safe behavior online	Students evaluate the ethical decision to share or not share AI-generated content, and discuss the harm each choice can cause
ISTE 1.3.b	Knowledge Constructor — evaluate accuracy, perspective, credibility	The SIFT protocol is the operational definition of this standard; every move in the walkthrough maps directly
CCSS.ELA-LITERACY.W.7.8	Gather relevant information, assess credibility of sources	The grade-7 verification task is designed as a direct evidence task for this anchor standard
CCSS.ELA-LITERACY.RI.7.7	Compare and contrast a written text to an audio or video version	Students compare fabricated text articles to deepfake video — same story, different format, same misinformation strategy
AI4K12 Big Idea #5	Societal Impact of AI	Discussion of how AI-generated misinformation affects communities, elections, and individual decisions grounds the lesson in real consequences

The downloadable AI Journalism and Fake News Lesson PDF includes all five standards listed here with activity-level citations — ready to attach to a lesson plan or curriculum map.

When Students Disagree: A 3-Step Debrief Protocol

Run any of the four case studies with a class and you will likely hit the same moment: half the room says “real,” the other half says “fake,” and both sides are confident. That moment is not a problem to smooth over — it is the most instructive ten minutes of the lesson. Here is a three-step debrief for it.

Step 1: Notice the disagreement before resolving it. Don’t tell students who is right. Instead, ask everyone to write down, on a sticky note or in their notes: “What specific feature of the video or article made you decide?” Give them ninety seconds. No sharing yet.

Step 2: Name the criteria together. Ask students to share their criteria. Write them on the board without commentary. A typical class generates six to ten criteria — and they’re often measuring different things. One student focused on visual quality: “The lips didn’t match the sound.” Another focused on source credibility: “The account only had twelve followers.” A third focused on content: “No real politician would say that.” When students see the range of criteria they applied, they realize they weren’t evaluating the same thing. That realization — that verification requires consistent criteria, not just gut response — is the central concept.

Step 3: Trace to primary source together. Return to the “T” in SIFT. Can the class find the original, unedited footage or the original article? For the Ireland and Canada cases, fact-checking organizations have documented the originals. Pull up the primary source as a class and watch whether the disagreement resolves. Sometimes it does. Sometimes the primary source is also contested — and that’s worth acknowledging too.

Facilitation note: the non-US political context of the Ireland and Canada cases significantly reduces emotional investment in who’s right. Literacy specialist Pamela Brunskill, cited in Chalkbeat’s October 2024 coverage of classroom misinformation work, put it directly: “Misinformation, if you believe it, changes the trajectory of your beliefs and your actions.” The debrief protocol makes that consequence concrete — students can trace exactly how a wrong call at Step 1 would have changed what they did next.

After the class debrief, students who need independent practice can work through the AI Fact-Check Activity (8 guided verification tasks, grades 6-8) — a good station or homework follow-up that keeps the SIFT moves active without requiring whole-class facilitation.

Already have students comfortable with debate? The AI debate activity pairs well as a next step — once students can verify a claim, asking them to argue a position about it requires a different and complementary skill set.

The skill this lesson builds is not the ability to spot a fake by instinct — it’s the habit of running a protocol before deciding. Students who finish this unit leave knowing what to do when something looks wrong, which is more durable than knowing what bad deepfakes looked like in 2025. The protocol is the point.

This post was drafted with AI assistance and human-finalized.