CS50 for Business - Lecture 4 - Approaching Artificial Intelligence

CS50 (Harvard University)
CS50 (Harvard University)Mar 18, 2026

Why It Matters

Understanding AI decision frameworks equips businesses to automate strategic choices, improve customer experiences, and stay ahead in data‑driven markets.

Key Takeaways

  • AI excels at games, handwriting, spam filtering, recommendations
  • Decision trees use questions to guide actions in simple games
  • Minimax assigns numeric values to outcomes for optimal play
  • Maximizer seeks highest score; minimizer seeks lowest score
  • AI algorithms must handle uncertainty beyond fixed-rule environments

Summary

In Lecture 4 of CS50 for Business, David Malan and Brian Yu introduce the fundamentals of artificial intelligence, outlining its core purpose, capabilities, and constraints. They frame AI as a set of techniques that enable computers to interpret inputs—such as game states, handwritten digits, or email content—and produce intelligent outputs like moves, classifications, or recommendations.

The presenters walk through classic AI use cases, from early game‑playing research to modern recommendation engines and large‑language models. They illustrate decision‑making with simple pseudo‑code for a breakout‑style game, then deepen the discussion with tic‑tac‑toe strategies that ask binary questions (e.g., “Can X win on this move?”). This leads to the introduction of the Minimax algorithm, which converts game outcomes into numeric scores (‑1 for O win, 0 for tie, +1 for X win) and recursively evaluates all possible moves to select the optimal action for both maximizing and minimizing players.

Key examples include the step‑by‑step reasoning: “If the ball is to the left of the paddle, move left; otherwise, if it’s to the right, move right; else stay still.” In tic‑tac‑toe, the algorithm evaluates potential O moves, predicts X’s response, and chooses the branch with the lowest score for O, demonstrating how optimal play leads to a tie when both sides act rationally.

The lecture underscores that AI is not about hard‑coding every decision but providing frameworks—like decision trees and Minimax—that let machines autonomously resolve complex problems. For businesses, mastering these concepts translates to building smarter automation, predictive analytics, and competitive recommendation systems that can adapt to dynamic environments.

Original Description

This is Lecture 4 of CS50 for Business on Approaching Artificial Intelligence. Explore the fundamentals of artificial intelligence, including game playing algorithms, reinforcement learning, neural networks, and natural language processing.
To take this course for a certificate, register at cs50.edx.org/business.
***
This is CS50, Harvard University's introduction to the intellectual enterprises of computer science and the art of programming.
***
HOW TO SUBSCRIBE
HOW TO TAKE CS50
Harvard Extension School: https://cs50.harvard.edu/extension
Harvard Summer School: https://cs50.harvard.edu/summer
HOW TO JOIN CS50 COMMUNITIES
HOW TO FOLLOW DAVID J. MALAN
***
CS50 SHOP
***
LICENSE
CC BY-NC-SA 4.0
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International Public License
David J. Malan
malan@harvard.edu

Comments

Want to join the conversation?

Loading comments...