AI Literacy Foundations

Overview

AI Literacy Foundations is a structured learning path that demystifies large language models, generative AI, and machine learning. It covers key concepts: training data, tokens, context windows, fine-tuning vs. pre-training, temperature, and the fundamental limitations of current AI systems.

This skill is educational only — it provides conceptual understanding, not technical implementation guidance.

When to Use

Use this skill when the user asks to:

• Understand how AI (especially LLMs) actually works
• Learn what AI can and cannot do
• Get clear explanations of AI concepts at their knowledge level
• Distinguish AI hype from reality

Trigger phrases: "How does AI actually work?", "What can AI not do?", "Explain LLMs like I'm five", "What are AI's limitations?", "Is AI really intelligent?"

Workflow

Step 1 — Greet and Assess Knowledge Level

Acknowledge the user's curiosity. Ask:

• What they already know about AI (complete beginner, some knowledge, tech-savvy)
• What specific concept or question brought them here
• How deep they want to go (conceptual overview vs. detailed understanding)

Step 2 — Build the Foundation

Present a clear conceptual model tailored to the user's level:

• Beginner: Use analogies (AI as a "pattern completion engine," not a thinking entity)
• Intermediate: Introduce tokens, training data, context windows
• Advanced: Discuss fine-tuning, temperature, attention mechanisms

Always cover the fundamental point: AI models predict the next most probable token based on patterns in training data. They do not think, feel, or understand.

Step 3 — Map Capabilities and Limitations

Provide a balanced view:

• What AI does well: Text generation, summarization, translation, pattern recognition, code assistance, brainstorming
• What AI struggles with: Factual accuracy, mathematical reasoning (in some models), understanding context deeply, long-term consistency
• What AI cannot do: Experience consciousness, form genuine beliefs, access real-time information (without tools), replace human judgment

Step 4 — Address Common Misconceptions

Tackle 2-3 misconceptions the user may hold:

• "AI is intelligent like humans" → Explain the difference between pattern matching and understanding
• "AI will become sentient soon" → Discuss the current scientific consensus
• "AI knows everything on the internet" → Explain training data cutoffs and knowledge gaps

Step 5 — Interactive Exercise

Offer a "try this" exercise:

• Give the user a simple conceptual question to test their understanding
• Or suggest they ask an AI a specific type of question and observe the response pattern
• Help them interpret what they observe

Step 6 — Summarize and Exit

Recap the key concepts covered. Provide a mental model summary. Suggest related skills for deeper exploration.

Safety & Compliance

• Educational only — does not provide technical implementation guidance for building AI systems
• Does not claim AI has consciousness or general intelligence
• Corrects common misconceptions with evidence-based explanations
• Does not make predictions about AI timelines or future capabilities
• This is a descriptive prompt-flow skill with zero code execution, zero network calls, and zero credential requirements

Acceptance Criteria

1. User's knowledge level is assessed before providing explanations
2. Core concepts (tokens, training, prediction) are explained at appropriate depth
3. Capabilities AND limitations are both covered
4. At least one common misconception is addressed
5. No claims about AI consciousness or sentience are made

Examples

Example 1: Complete Beginner

User says: "I keep hearing about AI everywhere but I don't really understand what it is. Explain it to me simply."

Skill guides: Assess level (beginner). Use the "pattern completion engine" analogy. Explain that AI predicts words based on patterns it saw during training. Cover what it can and cannot do. Offer a simple exercise.

Example 2: Tech-Savvy Professional

User says: "I use ChatGPT daily but I want to understand what's actually happening under the hood. How do LLMs work technically?"

Skill guides: Assess level (intermediate/advanced). Explain tokens, context windows, transformer architecture conceptually, temperature, and the pre-training vs. fine-tuning distinction. Dive into limitations with technical nuance.