AIP Student Series · Chapter 1 of 11 · Read This First

AI Literacy: What You Actually Need to Know

The chapter most students skip — and most students later wish they had read first

How AI Actually WorksThe Hallucination ProblemSchool Policy Rules

The Most Important Thing to Understand

AI Doesn't Know If It's Right. It Just Sounds That Way.

ChatGPT, Claude, Grok, Gemini — these tools were trained on an enormous amount of text. They learned patterns: how sentences are structured, how arguments develop, how different types of writing are organized. When you ask a question, the AI generates a response by predicting what text would make sense given your input.

Here is what most students miss: the AI does not know whether what it is saying is true. It knows what text is statistically plausible. A plausible-sounding response and an accurate response are not the same thing. The AI will write a wrong answer in exactly the same smooth, confident tone it uses when the answer is right. You cannot tell the difference by reading it. You can only tell by checking.

The hallucination problem: AI generates false information with apparent confidence. For students, this shows up in two dangerous ways: fake citations (sources that do not exist but look real) and wrong facts (dates off by a decade, statistics that are close but inaccurate, scientific findings that are mischaracterized). Students have faced academic integrity investigations not because they intended to cheat — but because they trusted AI output without verifying it.

What Each Platform Does Well

ChatGPT (OpenAI). Most widely used. Strong for brainstorming, drafting, concept explanation, study help. Free version handles most high school tasks well.

Claude (Anthropic). Particularly good at long-form analysis and careful reasoning. Strong for complex essays and research questions. Known for being straightforward about what it does not know.

Grok (xAI). Has real-time access to information. Useful when you need recent events or current news. Verification still required — real-time access does not mean guaranteed accuracy.

Gemini (Google). Integrates with Google Docs, Drive, and Gmail. Useful if your school uses Google Workspace.

The one habit that changes everything: Every specific fact, statistic, or citation that an AI generates must be verified against a real source before it appears in any school work. Every single one. This is not optional. The habit takes thirty seconds per fact. Not building it can cost you a lot more.

The Three Questions Before Every Use

Is this use permitted by my school's policy and my teacher's instructions? If you are not sure — find out. Do not assume.
Does this use help me learn, or does it replace the learning? Using AI to understand a concept helps. Having AI produce the work doesn't.
Would I be comfortable if my teacher could see exactly what I did? If no — that is important information.

Ready-to-Use Prompts

Copy these into ChatGPT, Claude, or any AI tool. Adapt them for your specific situation.

Verify AI Research Output

I asked an AI to help me find sources for a paper on [topic]. Here is what it gave me: [paste the citations]. I need to verify these are real. Please help me: (1) identify which of these citations look most suspicious and why, (2) tell me what I should search for to confirm each one exists, and (3) explain what to do if a source I cannot verify shows up in my notes.

Understand Your School's AI Policy

My school's AI policy says [describe what it says, or 'I'm not sure what it says']. I am working on [describe the assignment]. Please help me: (1) understand what this policy allows and does not allow for my specific situation, (2) identify any aspects of my planned AI use that might be in a gray area, and (3) suggest what I should ask my teacher to clarify before I proceed.

Chapter Quiz

AI Literacy: What You Actually Need to Know

5 questions — no limit on attempts.