Learning Objectives
By the end of this topic, you should be able to:
- Define and provide examples for: agent, sensor, actuator, procedural knowledge, declarative knowledge, strong AI, weak AI.
- Identify examples of procedural and declarative knowledge.
- Distinguish between reflex behavior, goal-directed behavior, and learning behavior in an AI agent, and give an example of each.
- Explain what the Turing Test was designed to measure, why producing human-like output is not sufficient evidence of intelligence, and how this limitation applies to modern AI systems.
Learning Activities
To help you meet the learning objectives, we have prepared three readings. They build on each other, so please complete them in order.
Readings
- Reading 1 - What AI Actually Is — the definition of the field, what an agent is, and why AI is already everywhere you look — most of it invisible
- Reading 2 - How AI Agents "Think" — reflex, goal-directed, and learning behavior; procedural vs. declarative knowledge; and the difference between AI built to perform and AI built to understand
- Reading 3 - The Turing Test and What It Reveals — the test, the ELIZA story, why the benchmark faded, and what it tells us about intelligence, appearance, and the systems we are building today
These readings intentionally build on each other, so please complete them in order.
Checking for Understanding
Review the Learning Objectives at the top of this page. The questions below will help you check your understanding before moving on to Topic 6B.
Agents and Behavior
- A modern car has sensors that detect whether the driver is drifting out of a lane and automatically corrects the steering. Is this reflex behavior, goal-directed behavior, or learning behavior? Explain your reasoning.
- A chess-playing program evaluates millions of possible move sequences and selects the one most likely to lead to a win. What kind of behavior is this? What distinguishes it from reflex behavior?
- A spam filter starts with a list of known bad words but over time adjusts its rules based on which emails users mark as spam. Describe this system in terms of agents, sensors, and actuators. At what point does it exhibit learning behavior?
- Explain the difference between procedural knowledge and declarative knowledge. Give one example of each from your own experience as a teacher — knowledge you have that fits each category.
Strong AI, Weak AI, and the Turing Test
- In your own words, what is the difference between strong AI and weak AI? Which category does a spam filter belong to? Which category does a chess-playing program belong to?
- Describe the Turing Test. What was it designed to measure, and why did Turing believe it was a useful measure of intelligence?
- What did the ELIZA program actually do when a user typed a sentence to it? Why did so many people believe it understood them? What does that tell us about the Turing Test as a benchmark?
- A student argues: "ChatGPT must be intelligent because I can have a full conversation with it and it always makes sense." Based on what you learned in the ELIZA reading, how would you respond?
It is completely fine to revisit the readings as you work through these questions.
Extend Your Learning
These optional topics go beyond the core learning goals for this topic but are rich avenues for deeper understanding.
-
Alan Turing's original 1950 paper
- “Computing Machinery and Intelligence” is surprisingly readable and raises philosophical questions that remain unresolved today. Turing himself anticipated most of the objections that people still raise against AI.
-
Joseph Weizenbaum's response to ELIZA
- Weizenbaum was so disturbed by how people responded to his program that he wrote a book — Computer Power and Human Reason (1976) — arguing for limits on AI development. His concerns feel remarkably current.
-
The Chinese Room argument
- Philosopher John Searle's famous thought experiment argues that a system can manipulate symbols perfectly without understanding any of them — a direct challenge to the idea that passing the Turing Test demonstrates intelligence.
-
The history of AI winters
- The field of AI has experienced two major periods of dramatically reduced funding and interest after early promises went unfulfilled. Understanding why helps put the current moment in perspective.