COMP 3200 - Intro to Artificial Intelligence - Lecture 01 - Course Introduction

Overview

Dave Churchill introduces COMP 3200 as a classical AI algorithms course focused on search, optimization, data structures, evolutionary algorithms, reinforcement learning, and neural networks—not on building large language models. He explains course logistics and assessment, prohibits AI-generated work on graded assignments while allowing AI as a limited reference, and argues that students need independent programming skills before relying on workplace AI tools. The lecture closes with competing ways to define intelligence, John McCarthy’s definition of AI, Alan Turing’s imitation game, and John Searle’s Chinese Room thought experiment.

Key takeaways

COMP 3200 centers on classical AI methods—search, optimization, data structures, evolutionary algorithms, reinforcement learning, and neural networks—rather than LLM construction.
The five assignments use basic JavaScript and HTML, may be completed solo or in pairs, and must not include AI-generated code; the midterm and final together make up 65% of the course grade.
Churchill permits AI as a limited reference for syntax or conceptual explanations, but students must independently solve graded problems and understand every line they submit.
Churchill’s workplace guidance is to build programming fluency before relying on AI: developers need enough understanding to evaluate, test, and explain generated code in interviews and on the job.
Alan Turing’s imitation game evaluates whether an interrogator can distinguish a computer from a human, while John Searle’s Chinese Room questions whether rule-following behavior demonstrates genuine understanding.
Students can turn assignment interfaces into portfolio evidence using screenshots or videos, but personalized assignment code must remain private and must not be posted publicly.

Chapters

0:00 Dave Churchill Introduces COMP 3200 and Lecture Access

Dave Churchill calls the course Intro to AI, though its official title is Algorithmic Techniques for Artificial Intelligence.
Lectures are recorded locally with OBS and a dedicated microphone, with room lecture capture as backup.
Churchill recommends attending and participating, noting that active students tend to do best and that personal contact helps him provide stronger references.

5:00 Brightspace, YouTube, Course Spreadsheet, and JavaScript Assignments

A Google spreadsheet links course materials and is updated after lectures; slides go to Brightspace (D2L), while edited lecture recordings are posted to YouTube.
Churchill maintains an archive of past course slides and lectures, but warns that roughly 10% of course content changes each year and current dates and materials take precedence.
Assignments arrive as personalized ZIP files and are submitted through Brightspace; students implement algorithms in basic JavaScript and HTML.

12:00 Assignment Extensions, AI Restrictions, Discord, and Exam Attendance

Each student receives one no-questions-asked 48-hour assignment extension; otherwise, late work loses 5% per hour, rounded up, and deadlines are 11:59 p.m.
AI-generated code from ChatGPT, Copilot, or other tools is prohibited on graded work; confirmed academic misconduct is referred to the department.
Discord participation is optional; students should use real-name nicknames and must not post assignment code or solutions.
Churchill plans in-person midterm and final exams, recommends attending class, and says scheduled lectures will be covered even if a class must be missed.

18:00 Classical AI Algorithms Define the Course Scope

The course emphasizes pre-LLM AI methods, especially search algorithms, optimization, data structures, and evolutionary algorithms.
Reinforcement learning and neural networks appear toward the end, but the course is not a guide to building LLMs.
Students practice representing problems with data structures, implementing algorithms from scratch, and weighing trade-offs such as memory use versus speed.

21:00 Russell and Norvig, Study Slides, and Lecture Materials

The first two-thirds of the course draws substantially from the Russell and Norvig AI textbook, which is recommended but not required.
Churchill designs PDF slides as study materials and says exam content is drawn directly from them, while lectures add demonstrations and explanations not always shown on slides.
PDFs may show only a video’s first frame; students should revisit the corresponding YouTube lecture to see embedded videos and animations.

23:00 Five Assignments, In-Person Exams, and JavaScript Tools

The course has five assignments, and students may work alone or in groups of up to two, changing partners between assignments.
The midterm and final together account for 65% of the course grade; Churchill describes the exams as comprehensive but based on examples and algorithms covered in class.
Assignments use basic JavaScript concepts—functions, classes, objects, loops, and arrays—and provide visual interfaces for students to implement backend algorithms.
Any editor is acceptable, though Churchill recommends VS Code; JavaScript instruction is limited because COMP 3200 assesses algorithms, not the language itself.

28:00 Personalized Assignment Files and Academic Integrity Enforcement

Churchill says assignment ZIP files contain students’ identifying information in plain-text metadata and hidden text to discourage public sharing.
Publicly posting assignment code is forbidden, and GitHub repositories used with a partner must be private.
Churchill says staff compare submissions against prior assignments and investigate suspected misconduct using evidence beyond intuition.
Students should seek help in office hours when stuck; Churchill says he can often recognize recurring assignment bugs and help resolve them.

33:00 Using AI as a Reference Instead of an Assignment Solver

Churchill argues that AI can solve many take-home assignments, making independent in-person exams important for evaluating what students know themselves.
He permits using AI as a reference for syntax or explanations—for example, finding JavaScript array-sorting syntax—but not to generate assignment solutions.
His rule of thumb is to think through problems yourself when possible, consult tools for facts or syntax you could not derive by reasoning, and never submit code you do not understand.
Churchill frames the course’s purpose as building students’ ability to solve problems, not merely producing working assignment files.

40:00 Programming Fluency, AI at Work, and Hiring Interviews

Churchill expects graduates to use AI coding tools in industry, but argues they must first understand the code well enough to review and test it.
He warns that relying on AI to write unfamiliar algorithms can weaken programming ability, drawing on his own experience using AI extensively over the summer.
He compares AI-written assignments to using a forklift at the gym: a shortcut may hide missing skills that become apparent in a job interview.
His advice is to learn independently in class, then gain practical experience with AI tools in the workplace.

46:00 Tech Hiring Trends, Computer Science Enrollment, and Student Portfolios

Churchill describes pandemic-era tech hiring as an overexpansion followed by layoffs, while citing a roughly 20% increase in U.S. tech jobs over the prior 12 months and signs of improvement in St. John’s.
He says computer science enrollment surged during the pandemic and later fell; the department had no competitive major entry the previous year, and his COMP 4300 enrollment had dropped from a historical low of 53 students to 17.
Churchill cautions that a degree no longer guarantees a job and advises students not to assume that graduate school automatically improves employment prospects.
Students may use assignment screenshots and videos in a portfolio or résumé, but may not publish the assignment code online.

55:00 Defining Intelligence and John McCarthy’s Artificial Intelligence

Churchill distinguishes defining intelligence by outcomes—such as a correct chess move or useful recommendation—from defining it by a particular biological method.
He offers learning, understanding, and problem-solving as a broad working description of intelligence, while defining AI as a machine or program that appears intelligent in a domain.
John McCarthy described AI as the science and engineering of making intelligent machines, especially computer programs, without requiring biologically observable methods.
Churchill notes that AI spans machine learning, natural-language processing, speech, expert systems, planning, optimization, robotics, and computer vision.

1:00:00 Alan Turing’s Imitation Game and the Turing Test

Alan Turing proposed the imitation game as a practical alternative to settling abstract definitions of “machine” and “think.”
In the test, an interrogator communicates through a barrier with a human and a computer and tries to determine which is which.
Churchill says computers can outperform people in particular domains, but argues that human-indistinguishable behavior has not been achieved universally.
He points to The Imitation Game as a popular account of Turing’s work, including codebreaking during World War II.

1:04:00 John Searle’s Chinese Room Challenges Machine Understanding

John Searle’s 1980 Chinese Room thought experiment imagines a person using a rulebook to translate unfamiliar-language symbols without understanding the language.
The room can produce convincing responses, yet Searle argues that manipulating symbols by rules does not establish semantic understanding.
Churchill connects the argument to computers as deterministic systems that transform inputs according to implemented rules, including neural networks built on linear algebra and statistics.
The example challenges whether a system’s successful behavior, including passing a Turing-style test, proves that it understands what it is doing.

Summary, takeaways, and chapters were generated by AI from the video's transcript and may contain errors. The video belongs to its creator, Dave Churchill.