Explain It to Grandma

Overview

Families, students, and board members are forming opinions about AI in schools right now, often from headlines. The educators who can explain it plainly and honestly, without hype or fear, shape that conversation. In this session participants pair an audience card with a concept card, draft a 90-second explanation, and then test it on a colleague who plays that audience and pushes back with the audience's real questions. Explanations get scored for plain language and for accuracy, because simple is only useful if it's also true.

Objectives

• Participants will explain one AI concept (such as how chatbots generate text, hallucination, bias, or data privacy) in plain language suited to a specific audience.
• Participants will keep their explanation technically accurate while simplifying it, including at least one honest limitation.
• Participants will revise their explanation based on role-play pushback and name what they changed.

Step by step

1. 10–4 min

   Hook

   The jargon wall

   Read this aloud in your most serious voice: "Large language models use transformer architectures to perform next-token prediction over high-dimensional embeddings trained on web-scale corpora." Ask: "Hands up if a parent at pickup would find that reassuring." Then: "It's accurate. It's also useless to most people. Today we aim for explanations that are both plain and true."

   Facilitator noteStart a "Words that confused" list right away: model, token, algorithm, training, prompt.

2. 24–9 min

   Model

   Watch one, then critique it

   Perform your 90-second explanation of Concept C1 ("What is a generative AI chatbot?") to a volunteer holding Audience card A1 (the grandparent). Ask the room to score it with Handout C. Then name your moves: an everyday comparison ("like a friend who's read a huge library and is very good at guessing what comes next"), one honest limit ("it can sound sure and be wrong"), and one action ("so we check important answers").

   Facilitator noteInvite critique of your comparison. Every analogy breaks somewhere. The library friend doesn't actually remember books word for word. Naming where your analogy breaks is a mark of accuracy.

3. 39–19 min

   Create

   Draw, draft, trim

   Each person draws one Audience card and one Concept card, then drafts on Handout B: the one sentence the audience must remember, an everyday comparison, one honest limit, one thing the audience can do, and a jargon swap list. Trim to 90 seconds when spoken. Trio partners can help, but each person owns their own draft.

   Facilitator noteIf someone draws a mismatch that feels impossible (bias for a second grader), that's the challenge. Offer: "If the AI mostly learned from pictures of one kind of family, what might it think a family looks like?"

4. 419–34 min

   Practice

   Role-play rounds

   In each trio, rotate three roles for three five-minute rounds: Explainer (delivers the 90 seconds), Audience (reads the Audience card, stays in character, and asks at least one of the card's pushback questions), and Checker (holds the Concept card back and Handout C, and scores plain language and accuracy separately). After each round the Checker gives one "keep" and one "fix." Explainers mark changes on Handout B.

   Facilitator noteWatch for the most common trade-off: simplifying into something false ("AI searches the internet for the answer"). The Checker's job is to catch that without killing the plain language.

5. 534–41 min

   Debrief

   Words that worked

   Each trio adds two phrases to "Words that worked" and one to "Words that confused." Ask the room: "Which audience was hardest, and why?" and "Where did someone simplify so much it became untrue?" Highlight explanations that included an honest limit. Those tend to build more trust than ones that only reassure.

   Facilitator noteLeaders often find the school board card hardest because trustees want a decision, not a concept. Point them to the leader five questions: vision, evidence, instruction, who is left out, how families will understand.

6. 641–45 min

   Transfer

   Where will you use it?

   Each person writes on the back of Handout B the real moment they'll use this explanation in the next month (a family newsletter, open house, a board update, the first day of a unit) and the one sentence they'll keep. Collect the "Words that worked" chart photo for a campus-wide plain-language glossary.

   Facilitator noteLibrarians and coaches can turn the best explanations into a short family FAQ; offer to co-edit it.

Adaptations

Campus and district leaders

Use only Audience cards A3 (board trustee), A5 (reporter), and A6 (privacy-worried parent). Add a requirement that each explanation ends with what the district is doing and how families can ask questions.

K–2 and 3–5 teachers

Replace the planner with a three-box version: "AI is like…", "But it can…", "So we always…" and practice on Audience card A2 (the second grader).

Families who speak languages other than English

Have explainers practice with an interpreter role (Audience card A7). Plain language translates better; idioms and jargon don't. Note which comparisons fall apart in translation.

Pairs well with

• Gen AI Literacy Breakouts (K–8)Student-facing activities that pair well with the explanations teachers draft here.
• TCEA ELEs (PDF)The full AI in Education indicators, including AI1.2 on explaining AI to stakeholders.