← AI Agents & Automation

⏱ About 10 min✦ 10 XP

Sense-Think-Act Game

You have learned so much about how AI agents work. You know about sensors. You know about thinking and deciding. You know about acting and checking. You know the sense-think-act loop. Now it is time to BE the agent. In this lesson, you and your classmates, friends, or family are going to play a game where each person takes a role inside a pretend AI agent — and you will run the loop together, live, with real decisions and real actions. This is how some of the best AI engineers in the world say they understand AI: by acting it out themselves!

The Roles in the Game

The Sense-Think-Act Game is built for three to six players. Each player takes a role. Here is what each role does. The Sensor (one or two players): Your job is to observe the world and report what you see. You are like the eyes and ears of the agent. When the round starts, you look at the current situation and tell the Thinker exactly what you observe. The Thinker (one player): Your job is to listen to what the Sensor reports, think about it, and decide what the agent should do. You are the brain. After you decide, you tell the Actor what action to take. The Actor (one player): Your job is to carry out the action the Thinker chose. You act it out — say it out loud or physically do it. The World (one or two players): Your job is to set up each situation and then react to what the Actor did. Did the action work? What happened next? You tell the group and set up the next round.

The Setup

You need: 3 to 6 players, a set of Situation Cards (you will make them in the activity below), and a simple scoring paper. Play five rounds. After each round, rotate roles so everyone gets to be the Thinker and the Actor at least once.

The Full Sense-Think-Act Game

  1. SETUP (5 minutes):
  2. Make eight Situation Cards. On each card, write a scenario an AI agent might face. Use these examples or make your own:
  3. Card 1: A delivery robot is rolling down a hallway and sees a puddle of water.
  4. Card 2: A home assistant hears someone say 'I am cold.'
  5. Card 3: A robot arm is told to pick up a glass but the glass has liquid in it.
  6. Card 4: A navigation agent plans a route but senses a road is closed.
  7. Card 5: A homework agent sees a student has an overdue assignment.
  8. Card 6: A robot cook is told to add salt to soup but the salt shaker is empty.
  9. Card 7: A music agent is playing a song and hears someone say 'turn it down.'
  10. Card 8: A robot cleaner senses its battery is at 5 percent.
  11. HOW TO PLAY:
  12. Assign roles: one Sensor, one Thinker, one Actor, one World. Remaining players can be extra Sensors or help the World.
  13. The World draws a Situation Card and reads it aloud.
  14. The Sensor studies the card and reports to the Thinker: what does the agent know right now? Be specific.
  15. The Thinker has 30 seconds to think quietly, then announces the decision: what action should the agent take?
  16. The Actor says the action out loud in a complete sentence, as if they ARE the agent doing it.
  17. The World decides: did it work? They describe what happened next. If it worked, the group earns 1 point. If it did not quite work, the group talks about a better action and earns half a point.
  18. Rotate roles after each round. Play all eight cards or until everyone has been the Thinker at least once.
  19. AFTER THE GAME:
  20. Tall about these questions together:
  21. Which role was hardest — Sensor, Thinker, or Actor? Why?
  22. Was there ever a situation where the Sensor missed something important?
  23. Was there a situation where the Thinker chose an action that did not quite work? What would have been better?
  24. How is this game like what a real AI agent does inside a computer thousands of times per second?

After playing the game, think about this: you just ran the sense-think-act loop yourself, as a team. Each loop only took about a minute. A real AI agent can run that same loop hundreds or thousands of times every single second. Imagine doing everything your Sensor, Thinker, and Actor did — but a thousand times faster, without stopping. That is what happens inside an AI agent while it is working. Every fraction of a second, it is sensing, thinking, and acting. Over and over.

You Are the Agent!

Playing this game gives you something special: you know what it FEELS LIKE to be an agent. You had to notice things carefully, make decisions with incomplete information, and take actions that might or might not work. That is exactly the challenge every AI agent faces!

Flashcards — click each card to reveal the answer

In the game, the Sensor says: 'The delivery robot sees a closed door.' The Thinker says: 'Press the door-open button.' Which step of the sense-think-act loop does the Thinker's decision represent?

After the Actor carries out an action, the World says it did not quite work. What should the group do?

← Back to AI Agents & Automation