Social Media Algorithm Lesson Plan for Middle School

Every student in the room can tell you their For You Page shows them exactly what they want to see. Not one of them can explain why. When a principal hands a teacher the “teach media literacy” mandate — usually attached to a curriculum-night slide, rarely attached to a lesson — the algorithmic feed is the obvious entry point and the hardest one to explain without materials. That is a literacy gap, not a discipline problem. Once students can see the loop that runs their feed, the magic dissolves and the thinking starts.

A social media algorithm lesson plan for middle school teaches how recommendation feeds actually work — the mechanical loop first, the ethics second. This post gives you a plain definition, a 45-minute plan with a minute-by-minute table, an unplugged card simulation that needs zero devices, a standards crosswalk, and three subject-area entry points. A real teacher should be able to run it Monday.

What is a social media algorithm? (start class here)

A social media algorithm is a set of rules a platform runs to decide which content appears in a student’s feed and in what order. The loop has four steps: Signal → Score → Serve → Repeat.

Signal — Every interaction a user takes is logged: how long they watched a clip, whether they liked it, whether they shared it, or whether they scrolled past in under two seconds. These are signals.

Score — The platform’s model assigns a predicted-engagement score to every piece of content based on past signals from that user and from millions of users with similar patterns.

Serve — The highest-scoring content appears next. Not the newest. Not the most accurate. The content the model predicts will keep this specific user watching.

Repeat — The user’s next action creates new signals, which update the scores, which shape the next serve.

Here is the worked example to put on the board. A student watches three skateboarding clips to the end (high watch-time signal). They like one and share it (engagement signal). They scroll past a cooking video in under two seconds (skip signal). The model now knows: skateboarding = high score; cooking = low score. The next serve is more skateboarding. And the serve after that. The feed narrows in real time around what keeps this student engaged — not what is most useful, most accurate, or most broad.

This is exactly what AI4K12 Big Idea #3 (Learning — computers learn patterns from data) describes at the middle-school level. The algorithm is not sentient. It learned a pattern from data, and it is running that pattern.

A 45-minute social media algorithm lesson plan, minute by minute

Minutes	What you do	Why
0–5	Warm-up: Ask students “Name three things on your For You Page right now. Why do you think they’re there?” Cold call two or three answers, write them on the board.	Activates prior knowledge; surfaces misconceptions before instruction.
5–15	Definition lab: Project the Signal → Score → Serve → Repeat loop. Walk through each step using the skateboarding worked example. Students copy the loop into notes with one personal example from their own feed.	Builds the foundational mental model before any hands-on activity.
15–30	Unplugged simulation: Run the card role-play (full description below). One student plays “the algorithm,” others hold signal cards, the class enacts two rounds of the ranking loop.	Kinesthetic encoding — students experience the logic rather than just hear it.
30–40	Debrief discussion: “What happened to content with zero likes and low watch time? What would happen if the algorithm only had two types of content to choose from?” Introduce the phrase filter bubble as a natural consequence of the loop.	Bridges mechanics to critical literacy — the ethics section starts here.
40–45	Exit ticket: Students write one sentence: “The algorithm serves ___ because ___.” Collect before dismissal.	Formative check on whether students can apply Signal → Score → Serve logic independently.

A typical class finishes the exit ticket with two or three minutes to spare. Use that time to preview the next session: what happens when the content the algorithm scores highest is also the most emotionally charged.

The unplugged algorithm simulation (no devices needed)

This card-based role-play is the “how does the TikTok algorithm work” question made physical — students stop receiving an explanation and start running one.

Setup (under five minutes). Print or write signal cards on index cards. Make four types: Watch (student watched the full clip), Like (student tapped the heart), Share (student sent it to a friend), and Skip (student scrolled past in under two seconds). Assign one student the role of “the algorithm.” Everyone else holds cards.

Round one. Write three pieces of “content” on the board: a skateboarding clip, a cooking video, a news clip. Each student silently assigns one signal card to each piece of content based on what they imagine they’d actually do. The algorithm student collects the cards and tallies the scores: Watch = 4 points, Like = 3 points, Share = 5 points, Skip = 0 points. The highest-scoring content is “served” next.

Round two. Remove the lowest-scoring content from the board entirely. Add two new pieces of content similar to the winner of round one. Repeat the scoring. Students notice the board is already narrowing toward one category.

Debrief prompt. “What would the board look like after ten rounds? Fifty rounds?” Students name the filter bubble classroom activity concept on their own — the term lands harder when they derived the result themselves.

This directly addresses ISTE 1.5.d — students use algorithmic thinking by enacting the decision rules of a real system. No screens required.

For a ready-made version with printed signal cards, the student scoring sheet, and the debrief discussion guide: AI Intro Lesson Grade 7 — Feed Algorithm + Definition Lab.

Teaching filter bubbles and how feeds maximize engagement

The mechanics lesson sets up a natural question: if the algorithm always serves what keeps you watching, what does the feed stop showing you?

A filter bubble is what happens when the Signal → Score → Serve loop repeats long enough that the feed narrows to a small slice of the available content landscape. Students who watch politically charged content get more. Students who watch anxiety-provoking content get more. The algorithm did not intend to narrow anyone’s view of the world — it was scoring engagement, and emotionally charged content tends to score high.

“Maximizing engagement” is the plain-words name for what the algorithm is actually doing: it adjusts what it serves to increase the total time a user spends on the platform. This is a business objective. Students benefit from knowing that.

Bias enters the loop at two points. First, at the training-data level: if the data the model learned from over-represents certain voices and under-represents others, those patterns carry into the scores. Second, at the serve level: content that provokes strong reactions — anger, fear, excitement — tends to generate more signals per second, so it scores higher regardless of accuracy or source.

The critical-literacy standards that anchor this section are ISTE 1.3.b (evaluate the accuracy, perspective, credibility and relevance of information, media, data or other resources) and CCSS.ELA-LITERACY.RI.8.6 (determine an author’s point of view or purpose in a text and analyze how the author acknowledges and responds to conflicting evidence). The algorithm is an author in the informational-text sense: it makes choices about what information reaches a reader, and those choices reflect a purpose.

States keep adding media literacy to standards documents. What most of them do not add is the materials to teach it. The loop above — mechanics before ethics — is the specific move that closes that gap. Students who understand why the feed narrows are in a fundamentally different position than students who are simply told it is a problem.

For the full paired lesson — a social media gallery walk and a bias decoder activity in which students audit five different feeds for pattern and source: AI Everyday Life Lesson Grade 7 — Social Media Gallery Walk + Bias Decoder.

Standards crosswalk: what this lesson covers

Activity	ISTE	CCSS	AI4K12
Signal → Score → Serve → Repeat definition lab	ISTE 1.2.a (digital identity — understanding what data the platform collects)	CCSS.ELA-LITERACY.RI.8.6	AI4K12 Big Idea #3
Unplugged algorithm simulation (card role-play)	ISTE 1.5.d (use algorithmic thinking)	—	AI4K12 Big Idea #3
Filter bubble debrief discussion	ISTE 1.3.b (evaluate accuracy, perspective, credibility, relevance)	CCSS.ELA-LITERACY.RI.8.6	AI4K12 Big Idea #5
Source audit / bias decoder extension	ISTE 1.3.b	CCSS.ELA-LITERACY.W.7.8	AI4K12 Big Idea #5
Exit ticket — “The algorithm serves ___ because ___“	ISTE 1.2.a	—	AI4K12 Big Idea #3

Standards cited by anchor code. ISTE Student Standards: iste.org/iste-standards. ISTE is a registered trademark of the International Society for Technology in Education. These resources are not affiliated with or endorsed by ISTE.

How to introduce this in your subject

ELA. Open with an informational-text move: the algorithm is an author. Ask students to identify its purpose (keep users on the platform), its intended audience (the specific user’s past behavior), and its perspective (engagement over accuracy). That is an RI.8.6 analysis framing ELA teachers already know how to run. Pair it with the AI in everyday life lesson for a two-period arc.

Social studies. The filter bubble debrief lands cleanly inside a civics unit on media and democracy. Students compare two simulated feeds — one that narrowed after partisan signals, one that broadened — and discuss what it means for an informed electorate when every citizen’s information environment is personalized. The algorithm bias lesson ELA middle school unplugged simulation works in a social studies period with zero adaptation.

Advisory / homeroom. This is the lightest-lift entry point. Run only the warm-up and the exit ticket as a 15-minute check-in. Students name the loop. You name it back. No simulation required. Follow up the next week with the filter bubble discussion as a five-minute journal prompt: “What has your feed stopped showing you?” For teachers who want to extend into the machine-learning mechanics underneath the feed, unplugged machine learning activities gives the CS-angle companion lesson.

The literacy gap here is real, and it is fillable. The loop is teachable. The materials exist. You do not have to build this from a blank document on a Sunday when the week is already stacked.

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