LEARNING GOAL
Learning to differentiate a human being from an "intelligent" machine
PEDAGOGICAL INTENTION
Identify cognitive differences
Target SKILLS
MATERIALS
DETAILED PROCEDURE
PROGRESS INDICATORS
Computer/projector
List of AI benchmark performances
List of simple tasks AI fails at
Present : AI can beat the world chess champion, diagnose cancer from scans at radiologist level, write publishable code, and pass the bar exam.
Step 1
Then: AI cannot reliably tie a shoelace (robotics), understand that a glass of water placed on a tilted surface will spill (physical intuition), or know that it doesn't know something (metacognition).
Step 2
Student task: map these capabilities on a 2x2 grid - narrow/broad x symbolic/embodied.
Step 3
Discussion: 'What does this tell us about what 'intelligence' really means?' 5. Synthesis: distinction between task performance (narrow, optimized) and general understanding (contextual, embodied, metacognitive). Students write a definition of 'genuine understanding' that an AI currently cannot meet.
Step 4
Step 5
Step 6
Step 7
Step 8
Step 9
Step 10
Student produces a written definition of 'understanding' that includes at least three dimensions absent from current AI systems - e.g., embodiment, metacognition, transfer learning across radically different domains, common-sense reasoning.