Teaching "AI and Jobs" to Hacker mode (ages 11–14)

Hook & Warm-Up

Hacker mode students are old enough to have opinions about this topic already, often strong ones, formed from social media rather than any careful analysis. Use that as an opening, not a problem to correct immediately.

Say: "AI is changing the job market. Routine tasks get automated. New roles are emerging. Let's explore what careers look like in an AI-powered world." Then ask directly: "Show of hands — who's seen a video or post online claiming AI is going to take basically everyone's job?" Expect most hands to go up. Follow with: "Now — who's seen an actual, careful breakdown of which specific jobs, and which specific parts of jobs, are actually at risk?" Expect far fewer hands. Land the point: "That gap — between confident claims and careful analysis — is exactly what we're going to close today."

This age band responds well to being treated as capable of nuance rather than needing a simple reassurance. It's fine, and more honest, to say up front: "The truth is somewhere between 'nothing will change' and 'everything is over' — and by the end of this lesson you should be able to explain, with actual reasons, why some jobs and tasks are more exposed to automation than others."

If a student pushes back with a specific claim they've picked up online — "I saw a video saying [job] is completely dead in five years" — resist the urge to simply say they're wrong. Instead, treat it as a testable claim: "What would that video need to show us to actually prove that? Is it showing us a whole job disappearing, or one task inside that job being automated?" That question is the entire analytical move this lesson is trying to teach, and using a student's own example is far more effective than a generic one.

It's also worth naming, once, why this topic generates so much confident-sounding content online in the first place: strong claims ("everyone will lose their job" or, just as overconfident, "AI can never really replace anyone") get far more attention than careful, hedged, task-by-task analysis. That doesn't make either extreme true — it just explains why the loudest voices on this topic tend to be the least careful ones, which is a useful thing for this age group to notice about online information generally, not just about this specific topic.

Main Activity

Have students work through the three scenes at their own pace first, then run the debrief focused on the analytical skill underneath each example — sorting a task by whether it's routine and pattern-based (higher automation risk) or non-routine and relationship/judgment-based (lower automation risk).

Scene 1 — Jobs Changing. The lesson's four examples: a doctor whose diagnostic work is increasingly AI-assisted ("AI won't replace doctors — but doctors who use AI will replace doctors who don't"); a lawyer whose document review is now largely automatable (AI reviewing 10,000 documents in minutes, but unable to argue in court or comfort a client); a teacher, whose core work — mentoring, relationship-building — resists automation almost entirely; and a designer, where AI can generate options fast but a human still makes the judgment call about what actually connects emotionally. Push the class to articulate why: in every case, the automated part is the routine, high-volume, pattern-matching slice of the job (scanning images, searching documents, generating draft options), and the resistant part is judgment, trust, persuasion, or emotional connection — things that don't reduce to a repeatable pattern in the same way.

The legal-document-review example is worth spending an extra moment on with this age group, since it's a genuine, current case of the pattern the lesson is teaching: document review and similar high-volume back-office work is exactly the kind of large-scale, business-process outsourcing work that has historically employed a significant number of Filipino workers, which makes this less of an abstract example and more of a live question worth taking seriously — without overstating it into a prediction about any specific company or role.

Scene 2 — New Jobs. Prompt engineering, AI ethics officer, and AI trainer are all real, current job categories worth discussing honestly. Prompt engineering is the skill of getting reliable, high-quality output from an AI system through careful instruction and iteration — it exists as a distinct skill because these systems don't automatically produce your intended result on the first try. AI ethics officers (and similar "responsible AI" roles) exist inside companies that build or deploy AI, checking systems for fairness and safety before they reach users. AI trainers are the humans who label data and correct model mistakes — a large, mostly invisible workforce behind almost every AI product in use today. Contrast all three against plumbing, which remains hard to automate specifically because it happens in constantly-varying, physically cramped, real-world spaces that don't standardize the way digital tasks do.

Scene 3 — Future Skills. This is where to introduce the vocabulary the quiz uses. The "augmentation hypothesis" is the idea that AI more often enhances what a human worker can do than fully replaces the worker — the lesson's creativity, empathy, and adaptability skills are exactly the kind of human contribution that stays valuable alongside an AI tool rather than being made obsolete by it. It's worth being precise with this age group about why these three specific skills: creativity resists automation because generating a genuinely original idea isn't the same operation as recombining patterns from past examples; empathy resists it because it requires actually experiencing and responding to another person's emotional state, not simulating a plausible response to one; and adaptability resists it because AI systems are typically narrow — very good at the specific task they were built and trained for, and much weaker at handling a genuinely novel situation outside that scope.

After the scenes, run the in-app timed sorting game, where students classify tasks — prompt engineering, plumbing, AI ethics review, teaching kids, AI model training, creative thinking, AI collaboration, emotional support — as AI or not-AI under time pressure. Use the debrief to push past "is this AI or not" into "why": for each item, ask what makes it routine/pattern-based versus non-routine/judgment-based, which is the actual analytical skill this lesson is building.

Worth flagging honestly during this debrief: several items resist a clean sort, and that's the point rather than a flaw in the game. "AI model training" involves both a highly routine component (repeatedly labeling similar examples) and a judgment-heavy one (deciding what counts as a correct label in an ambiguous case). "Teaching kids" is overwhelmingly relational, but a teacher grading a stack of identical multiple-choice quizzes is doing something closer to a routine task. Encourage students who finish the timed game to go back through their answers and identify one item where they'd now argue for the opposite category — that second pass is where the real thinking happens.

Discussion

• "Pick a job you might want someday. Break it into three or four separate tasks it involves. Which of those tasks look more automatable, and which look more resistant? Does that change how you feel about the job's future?"
• "The augmentation hypothesis says AI usually enhances jobs rather than eliminating them. Can you think of a case where that's clearly true, and a case where it might not hold up?"
• "Prompt engineering didn't exist as a job ten years ago. What's a skill you have right now that might become a real job title in the next ten years?"
• "AI ethics officers exist because AI systems can cause real harm if nobody checks them. What's one kind of harm you think an AI system could cause if left completely unsupervised?"
• "If creativity, empathy, and adaptability are the most AI-resistant skills, what's one thing you could actually do this year to get better at one of them?"
• "Is there a job in the Philippines you can think of that's especially exposed to AI automation, or especially protected from it, because of something specific to the local context — language, cost of labor, or the kind of work available here?"

Quiz Walkthrough

This quiz assumes the class has engaged with the routine-vs-non-routine framing from the activity above. If a student answers correctly without being able to explain the underlying reasoning, treat that as a signal to revisit the relevant scene rather than moving straight to the next question — a memorized right answer without the reasoning behind it doesn't transfer to a new example the way this lesson is meant to.

Wrap-Up & Extension

Close with: "Routine thinking tasks face the highest risk of automation. Creative and people-focused roles are more resilient. The best strategy isn't competing with AI — it's building skills that complement it."

Extension activity (20–25 minutes): Have students pick a career they're currently interested in and research (using a search engine, if available) one real way AI is already being used in that field. Have each student prepare a short summary answering three questions: what task is AI doing in this field, is that task routine or non-routine by the framework from today's lesson, and what does that suggest about how the job itself — not just the task — is likely to change rather than disappear. Have several students present to the class.

A sharper follow-up for a class that's engaged: assign small groups one of the four Scene 1 jobs (doctor, lawyer, teacher, designer) and have them argue, using the task-based framework, for why that job is more or less exposed to automation than the others — then hold a short class debate or ranking exercise. The goal isn't a single correct ranking; it's practicing the habit of justifying a claim about automation with specific tasks rather than a vague feeling about the job as a whole.

If a stronger group finishes early, give them a harder version of the same exercise: ask them to find one task inside their assigned job that seems automatable today but that they'd expect to become non-automatable again in the future — or vice versa — and explain their reasoning. This pushes past a static "automatable / not automatable" sort into thinking about how the boundary itself moves over time as the technology changes, which is closer to how economists and labor researchers actually think about this question.