AI ARCADE
Four AI agents. Four worlds. Pick one and explore what it can do — and how to use it like a pro.
🎮 House Rules (read first!)
- Be honest — AI helps you learn, it doesn't do the work for you
- Cite it — say when you used AI
- Check it — AI can be confident and wrong
- Follow the rules — your school's AI policy + age limits
- Ask Mr. Mendoza if unsure
Gemini
The arcade twins — power-ups for research, images, video and code. Insert coin!
PLAY ▸ 🧸 OpenAI · BuddyChatGPT
Your cuddly study buddy — ask, brainstorm, practice, get unstuck. So soft.
PLAY ▸ 💿 Microsoft · Y2KCopilot
Glossy chrome helper inside the apps you already use. Loading the future…
PLAY ▸ 🎨 Anthropic · RemixClaude
The creative one — big ideas, careful thinking, writing & building. Spray it loud.
PLAY ▸AI can make you look smart today and worse tomorrow, or it can make you genuinely better. The difference is how you play it. Here is what the research actually shows.
🎯 Try first, then check
Do the thinking yourself, then ask AI to check it or coach you.
Students who leaned on an unguarded AI tutor scored 17% worse on a later test they took alone. A hints-only version erased the harm. (PNAS, 2025)💬 Hints, not answers
Tell it: coach me, do not hand me the answer. That one rule was the whole difference in the study on the left.
📚 Ground it in real sources
Use NotebookLM loaded with your own notes and readings, so it answers from real material instead of making things up.
86% accurate versus 39% for a plain chatbot on the same sources. Still not perfect, so check it.🏋️ Build the skill first
If AI does it before you can, you become a fragile expert: your work looks fine until the AI is gone and you cannot fix it.
Shown in a study of new coders who used AI from the start.🔍 Check its work
AI is confident and sometimes wrong. Open the sources it cites, and ask a second AI for anything that matters.
Why our class works this way
Why oral exams and drafts? A similarity score cannot catch AI writing, and AI detectors are so unreliable they falsely flagged over 60% of real essays by English-language learners as AI. So instead of a machine deciding, you show your drafts and explain your thinking out loud. That protects you from a false accusation and proves the learning is yours.
Why some work is device-free? Not to punish you. The tests that decide your future, board exams, the clinical floor, and college, measure what you can do unaided. Device-free work is where you prove the skill is yours. The rest is where you learn to use AI like a professional.
The truth about cheating: it is not new. About 60 to 70% of students admitted to it for over 15 years before AI existed, and AI did not raise that. What changed is that the shortcut got easier. But the person who lets AI do the learning is the person who cannot do the job later. Integrity here means protecting your own skill.
For teachers: the staff Agent Hub ▸ (staff sign-in required)