For most teachers, "AI" arrived as a single word with a specific face — ChatGPT, in late 2022 — and has expanded ever since to describe everything from spellcheck to artificial superintelligence. The categories on this page are not a technical primer for engineers. They are a working vocabulary for a teacher who has to decide what to allow in their classroom, answer the student who asks what AI actually is, and read a vendor's marketing without taking it at face value.

For school use, AI is best understood as software that performs tasks associated with human thinking: finding patterns, classifying information, making predictions, generating content, or responding to prompts. This definition is intentionally broad. A scheduling system, adaptive practice platform, predictive dashboard, chatbot, image generator, and speech-to-text tool can all involve AI, but they do not carry the same risks.

Generative AI is the subset that changed the public conversation because it creates new outputs. A teacher can ask for a lesson draft, a family letter, a rubric, an explanation, a song, a study guide, or a set of practice questions. That creative fluency is useful, but it also creates new concerns about accuracy, authorship, privacy, intellectual property, bias, and whether students are doing the thinking that learning requires.

The Examined Classroom stance is explicit: AI can support human judgment, but it does not understand students, hold educational values, or exercise professional responsibility. That distinction keeps every recommendation here aligned with the site's Terms of Use and with current school-system guidance.

This vocabulary also matters when students ask. A ninth grader who is told "AI is just fancy autocomplete" learns to dismiss it. A ninth grader who is told "AI can mean a classifier, a generator, a predictor, or a system that infers emotions from a face — each built differently and behaving differently" has the start of a real mental model. The categories below are the version teachers can use for both audiences.

Teachers often hear one word — AI — used for everything from spellcheck to artificial superintelligence. The categories below separate what is actually running in classrooms today from speculative claims about what might come next. The point is not to predict the future, but to give educators enough vocabulary to ask better questions about any tool a vendor puts in front of them.

Narrow AI remains the only fully realized category schools are likely to encounter day to day. Reactive systems complete fixed tasks. Limited-memory systems use previous data to inform current output. Generative systems create new artifacts. Emotion AI and neural-decoding examples show how quickly the ethical stakes rise when systems infer feelings, attention, mental states, or health signals.

The takeaway is not that schools should adopt every category. It is that the risk profile changes when a system moves from helping a teacher draft a worksheet to interpreting a student's emotion, behavior, disability status, or learning pathway.

Mustafa Suleyman, writing in The Coming Wave, names Artificial Capable Intelligence as a step before Artificial General Intelligence. The term is not widely adopted outside that book, but the underlying observation matters for school planning: increasingly agentic tools create different governance problems than a chatbot that waits for a prompt. Educators do not need a prediction timeline. They need to understand why the jump from suggestion to autonomous action changes the oversight question.

A capable system is one that can complete complex, multistep tasks with less supervision. In schools, the near-term version is not a science-fiction teacher replacement. It is a tool that can plan a unit, draft communications, analyze student data, generate interventions, and coordinate workflow. That is powerful enough to require boundaries, because it can quietly shift decisions from people to systems.

AGI and ASI remain contested, future-facing concepts. They matter for ethical imagination and policy planning, but the immediate work for teachers and leaders is more concrete: keep humans accountable for educational decisions, require transparent tool review, and make sure automation does not narrow student opportunity.

Policy & Ethics

Boundaries, detectors, and school decisions

AI Ethics

Frameworks for ethical action

Educator Scenarios

Pressure-test the policy