AI in Healthcare

Learning Objectives

This lesson is short by design — a 5–15 minute core loop of a hook, a set of "scenes" the student taps through, and a quiz — so keep expectations for one sitting modest. By the end, a student should be able to:

• Name at least three places AI already shows up in healthcare — reading X-rays and scans, discovering new medicines, and monitoring patients through wearables like smartwatches — rather than treating "AI in medicine" as one vague, futuristic idea.
• Tell AI-driven tools apart from ordinary medical tools that just look high-tech. The lesson deliberately pairs each AI example with a plain one — a stethoscope, a band-aid, a thermometer, a syringe — so the student practices asking "is a computer learning a pattern here, or is this just a well-designed physical object?"
• Explain in their own words what it means for AI to "detect" or "screen" something. An X-ray AI is not looking at a photo and knowing what cancer is the way a person does — it has been shown many thousands of labeled examples and has learned to recognize a pattern. A drug-discovery AI is not inventing medicine from imagination — it is narrowing millions of candidate chemical compounds down to a short list worth testing in a real lab.
• State clearly that AI works alongside doctors, not instead of them. Every age band's quiz ends on this point, and it is the single idea worth protecting if class time runs short: the tools in this lesson assist diagnosis and discovery, they do not replace a clinician's judgment, a physical exam, or the doctor-patient relationship.
• (Ages 11 and up) Name one real limitation of medical AI — that its accuracy depends on the data it was trained on, and it can perform worse on patients who look different from that training data — so the tools are understood as powerful but imperfect, not magic.

Teacher Background

You do not need a technical background to teach this lesson well. The core idea underneath every example in "AI in Healthcare" is the same one the rest of AIQ teaches: an AI system is a pattern-finder that got very good at one narrow job by studying a huge number of examples of that job done correctly. Everything in this lesson is a variation on that.

Diagnosis (reading images). A radiologist spends years learning to look at an X-ray or a retina photo and recognize what a tumor, a fracture, or diabetic eye damage looks like. An AI model can be shown a very large number of images that have already been labeled by doctors — "this one has cancer," "this one doesn't" — and it gradually learns which visual patterns tend to go with which label. Once trained, it can look at a brand-new image it has never seen and give a prediction. This is genuinely useful and, in some narrow, well-studied tasks, the published accuracy rates are very high — but it is pattern-matching on pixels, not understanding illness the way a doctor does. It also only works on the narrow task it was trained for; an AI trained to spot lung nodules cannot suddenly diagnose a skin condition.

Drug discovery (finding new medicines). Discovering a new medicine traditionally means testing enormous numbers of chemical compounds in a lab to see which ones might fight a disease — extremely slow, expensive work. AI can be used to narrow that search computationally first: a model trained on known chemistry and known drug behavior can rank millions of candidate compounds by how promising they look, so human scientists spend their lab time testing the most likely dozens instead of blindly working through millions. The AI does not "invent" a cure from nothing; it makes an educated shortlist. The compounds still have to be synthesized and tested in real labs, then in real clinical trials, before anyone can call them a medicine.

Patient monitoring (wearables). A smartwatch with AI-assisted heart-rhythm detection is doing something similar to the imaging example: it has learned what a normal heartbeat pattern looks like versus an irregular one, from many recorded examples, and it flags anomalies for the wearer to take to a doctor. It is a screening tool, not a diagnosis — a flagged irregularity is a prompt to get checked, not a medical verdict.

One thing worth being upfront about, especially with older students: these are real, documented uses of AI in healthcare — this is not speculative "AI of the future" content — but the specific numbers used as classroom hooks (accuracy percentages, compound counts) are illustrative of the kind of results published in this field, not live statistics you should treat as constantly current. Treat them as "here is the scale AI operates at in this domain," not as a number to be quizzed on precisely.

It also helps to have a one-sentence answer ready for the question a curious student will eventually ask: "How does the computer actually learn this?" In plain terms — the model is shown a large number of past examples where the right answer is already known (an X-ray a radiologist has already diagnosed, a compound already tested for a property), and an optimization process gradually adjusts millions of internal numeric settings inside the model until its predictions on those known examples get closer and closer to the correct answer. Once that process has run its course, the same adjusted settings are used to make a prediction on a new case nobody has labeled yet. Nothing about that process resembles reasoning, understanding, or intuition in the human sense — it is closer to a very elaborate, very well-tuned form of pattern-matching, which is exactly why its accuracy is tied so tightly to how much good example data it was shown and how representative that data was.

A second thing worth anticipating: students at every age level, but especially Hacker and Architect, may ask whether these tools are already used in Philippine hospitals specifically. The honest answer is that AI-assisted diagnostic imaging and similar tools are used in some hospitals internationally and, increasingly, in some private hospitals and diagnostic centers in the Philippines, but adoption varies a great deal by facility and region, and this lesson is describing the technology and its use in the field broadly rather than making a claim about any specific local hospital. If a student asks and you don't know the local specifics, that is a fair and honest answer — it also makes a good extension research question for older students.

Materials & Prep

Nothing beyond what AIQ normally needs: one device per student (or per small group) with a modern browser and an internet connection to open aiq.ph. No login is required to work through the lesson. No printouts, no lab equipment, and no advance setup — the lesson is self-contained inside the app. If you plan to run the discussion or wrap-up as a whole-class conversation rather than silent reading, budget an extra 5–10 minutes beyond the in-app content, and it helps to have the board or a shared screen free for jotting down student examples during the discussion.

If your classroom has only one shared device or a projector rather than one device per student, the lesson works fine delivered as a group read-along — open it on the shared screen, read each scene aloud, and use the age-specific activity scripts linked below as your delivery guide rather than relying on students clicking through it themselves. This actually works especially well for Explorer and Builder mode, where the teacher-led pacing described in each age script is closer to how the lesson is meant to land anyway. For Hacker and Architect mode, individual device access is more valuable since parts of the activity ask students to read and reason about the content at their own pace before the group debrief.

No specialized vocabulary needs to be pre-taught. Every technical term the lesson uses — X-ray, retina scan, antibiotic, gene therapy, smartwatch — is something students at every age band in this range have likely already encountered outside of class. The one piece of vocabulary worth previewing for Hacker and Architect classes, if you have a minute before starting, is "training data" — the set of past examples an AI model learns from — since both of those quizzes build directly on that idea.

Common Misconceptions

Pick your grade's script