CS50's Introduction to Artificial Intelligence with Python

You build the ideas yourself — search, logic, probability, machine learning, neural nets — implementing each in your own Python, which is what makes them stick. With over 1.6 million learners and Malan and Yu teaching, it is a well-regarded free AI course, and the project-driven format is a big part of why.

It is not a true beginner course despite the 'introductory' label: it expects prior Python from CS50x or equivalent, and it will not go as deep into the mathematical theory as a graduate course. If you cannot yet program, start with CS50x or CS50P first. It is free to audit with an optional paid HarvardX certificate, and the CS50 team keeps it current (as of 2026).

CS50's Introduction to Artificial Intelligence with Python is HarvardX's deep dive into the ideas that power modern AI. Over seven self-paced weeks you work through graph and adversarial search, knowledge representation and logical inference, probability and Bayesian networks, optimization and constraint satisfaction, machine learning, reinforcement learning, neural networks and natural language processing — implementing each in your own Python programs through hands-on projects.