CS-215 AI Literacy
Topic 06 · Slides studio + graded team lab

AI Media, Accessibility, and EvidenceBrief the job. Make what we made. Then explain an agent idea in the graded team lab.

A deck an AI makes is a draft. Thursday 10/15: a picture, slides and a video made with Claude Code, each the same way: see ours, learn how we made it, make one yourself, and score it. Thursday 10/22: the graded team lab, three teams with one topic each (loop engineering, graph engineering or self-improving agents), making slides, a video or an animation checked with the same five lenses. Tuesday 10/27: each team presents it.

Before class

Thursday 10/15: bring a pen and your laptop with Claude Code. Thursday 10/22: the graded team lab, one class: three teams, one topic each, set in class. Bring a laptop with Claude Code: your team builds on one laptop; by hand only if Claude Code fails on the day. No other AI tools. Before class, watch the 14-minute Chase AI video linked below. Read the five lab sources in class. Tuesday 10/27: each team presents its piece, 5 minutes plus 2 for questions.

By the end

Four outcomes you can demonstrate.

The focus is judgment students can show, not vocabulary they can merely repeat.

01 · FRAME

Describe a picture, a deck or an animation precisely enough for Claude Code to make it again (canvas, where, look, words, details and, for motion, when), and brief a team’s piece in five boxes.

02 · CHECK

Name a slide’s first failure with one of five lenses and a fact a reviewer can see.

03 · EXPLAIN

Explain loop engineering, graph engineering or self-improving agents in slides, a video or an animation, built from sources you name.

04 · DISCLOSE

Say who made what in one sentence, credit each source correctly, and keep every claim traceable to a minute, a section or a quote.

Class route

Two meetings with separate learning arcs.

Meeting 1 plans 80 taught minutes with no separate buffer; Meeting 2 is an 80-minute lab.

MEETING 1

Make What We Made: a Picture, Slides, a Video

Make three things we made with Claude Code, each the same way: see ours, learn how we made it, make one yourself, and score it against five features. A picture (our synthwave, then the Study Night poster), slides (our deck; five lenses and an invented quote to watch for; then two rounds) and a video (our animation and two videos on making video, then a try).

80 MIN
  1. LaunchThree things we made with Claude Code: a picture, slides and a video; one check before trusting an AI’s draft.
  2. Make what we made, 1 of 3: a pictureOur picture and how we made it (the five-part card); then the poster, in words only, scored out of five.
  3. Make what we made, 2 of 3: slidesOur slides; five lenses and an invented quote to watch for; then our deck, Round 1 in words until 0:50 and Round 2 with our pictures until 0:57, a score after each.
  4. CloseThe words that got you closest; the exit line; how the team lab on Thursday 10/22 works.
  5. Make what we made, 3 of 3: a videoOur animation, the videos, our animation in words only, how your instructor made a video of Oreo the cat, then our exact prompts for the poster, the slides and the animation.
Start Meeting 1
MEETING 2

Graded Team Lab: Explain an AI-Agent Idea

One class, 80 minutes; the instructor does not teach. Three teams, one topic each (loop engineering, graph engineering or self-improving agents): each team reads any of the five sources, makes slides, a video or an animation with Claude Code (by hand only if it fails), checks it with the five lenses from 10/15, and one member submits it with the team’s five-line record on eCampus by the end of class. Each team presents it on Tuesday 10/27.

80 MIN
  1. Teams, roles, readYour instructor names the teams and topics; split the roles; read any of the five sources.
  2. BriefFive boxes before any request.
  3. BuildSlides, a video or an animation, with Claude Code (by hand only if it fails).
  4. Check and fix; plan the talkThe five lenses from 10/15; fix the first failure; plan the 5-minute talk.
  5. Record and submitOne member submits the file itself and the team’s five-line record on eCampus by 7:25 PM.
Lab pageTopic 6 Team Lab: explain an AI-agent idea

Three teams, one topic each: make slides, a video or an animation from any of the five sources, submit it on eCampus by 7:25 PM with your team’s five-line record, and present it on Tuesday 10/27.

Open the team lab
Evidence

Topic 6 Team Lab: explain an AI-agent idea

One class, 80 minutes (Thursday 10/22), in three teams with one topic each: loop engineering, graph engineering or self-improving agents. Read any of the five sources, make slides, a video or an animation with Claude Code (by hand only if it fails), submit it on eCampus by the end of class, and present it on Tuesday 10/27. The lab replaces the packet.

What to submit

  • What you made, as the file itself: your slide file (PDF, PPTX, KEY, or HTML; zip a folder), or your video (MP4, MOV or WebM) or animation (HTML, GIF or MP4). Too big for eCampus? A link that opens without signing in. One member submits for the team.
  • Your team’s five-line record: team, topic and each member’s role; sources used, with the minute, section or quote behind each claim; tool, skill (or “no skill”) and first prompt; the first failure you found and your fix; one disclosure sentence with its credit line.
  • Graded out of 100: the piece 80 (the five lenses from 10/15 at 16 each: brief fit, composition, text + fact, access, and boundary + disclosure) and the talk on Tuesday 10/27, 20 (the idea, where a claim comes from, what your sources don’t show, and time).
  • Submit on eCampus by 7:25 PM, the end of class. Tuesday 10/27: present from the submitted file, without editing it, for 5 minutes plus 2 for questions.
  • Never include real people’s faces or names, logos or frames copied from a source, account screens, private details or full chat history.
Open assignment
References

Core sources and optional extensions.

The class deck contains the required explanation. These links support verification, deeper reading, or alternate access.

Topic 6 Targets: make what we made (Thursday 10/15)Our picture (the worked example), then the poster, the deck and the animation, made with Claude Code: the rounds, the scores and our exact prompts.Video, before Thursday 10/22: “Claude Now Does Video (FOR FREE) Thanks To JavaScript” (Chase AI, YouTube, 14 min)How Claude Code makes video, in four levels. Watch it before the team lab on Thursday 10/22.Topic 6 Team Lab (graded, Thursday 10/22)The lab page: the rules, the roles, the three topics, the five sources, what to submit, the talk on Tuesday 10/27 and how it is graded.Source A · video, 86 min · posted on X by @res1dualedge, 9/28/2026Credit the poster; who made it is not verified, whatever the post says.Source B · video, 112 min · posted on X by @res1dualedge, 9/27/2026Credit the poster; who made it is not verified, whatever the post says.Source C · article, about 1,700 words · posted on X by @hanakoxbt, 8/23/2026Credit the poster; who made it is not verified, whatever the post says.Source D · video, 27 min · posted on X by @1006_amit7481, 9/28/2026Credit the poster; who made it is not verified, whatever the post says.Source E · article, about 1,400 words · learnAIDoc wiki: “Harness / Loop / Graph Engineering”Credit the wiki page.The three AI decks (pre-written for class)Three AI-made decks, each with the request it was made from; Case C on 10/15 shows the last slide of Deck 3.Video: “Claude Can Now Make Videos – EVERYTHING You Need to Know” (Christian Peverelli, YouTube)Optional, for teams thinking of making a video on 10/22. The lab’s tool rules still apply: Claude Code only.Claude Academy: “Turn research into presentations”Optional, 15 minutes: turn sources into a talk by telling Claude the audience, the time and who presents. Its example uses Claude’s chat and Canva; the lab’s tool rules still apply: Claude Code only.Claude for Microsoft 365Optional, outside class: Claude working inside PowerPoint, Word and Excel. In class and in the lab, Claude Code only.Higgsfield: photo-style pictures through ClaudeOptional, outside class: photo-style pictures from image models, with your own free Higgsfield account (a few free renders, then paid credits). No real people’s faces or names. In class and in the lab, Claude Code only.Motion design with Claude Code · optional practiceOptional: the credited Opus 5.5 video examples, an English overview, a complete Study Sprint seed, and an original prompt with checks and a paper route. No new assignment, required account, or Topic 8 prerequisite. Claude Code only.