The deal, in writing.
CSCI 100 — Critical Questions in Computer Science: Prompt Engineering. First-year seminar · Fall 2026 · 4 credit hours · COLL 100 attribute. The syllabus of record lives on Blackboard; this page is the same content, kept in sync.
Key dates
AI use
Expected, not a shortcut around it. What is graded is your judgment — how you direct the system, how you check what it gives back, and whether you can tell the difference.
Chapter One Who and when
Instructors
The course runs in two sections. Both follow the same weekly topics, assignments, and grading — check which one you are registered for.
| Section | Instructor | Meets | Room |
|---|---|---|---|
| 06 CRN 14423 | Dr. Antonio Mastropaolo amastropaolo@wm.edu · coordinator | MWF 10:00–10:50 AM | ISC 2280 |
| 09 CRN 14522 | Dr. Trey Woodlief adwoodlief@wm.edu | TR 2:00–3:20 PM | ISC 2280 |
Contact your own section instructor first. Dr. Mastropaolo handles anything affecting the whole course. Dr. Mastropaolo holds office hours Mondays and Wednesdays, 4:00–5:30 PM, in ISC 2333; Dr. Woodlief posts his on Blackboard in the first week. Both instructors are available by appointment.
Course calendar
| Wk | Topic | Milestones |
|---|---|---|
| PART ONE — LEARNING THE TECHNIQUE | ||
| 1 | What is AI, and what a prompt is | First day of classes. Diagnostic survey. |
| 2 | Anatomy of a prompt; zero-shot and few-shot | Last day to add/drop (Fri Sep 4). |
| 3 | In-context prompting and personas | Labor Day Mon Sep 7 — no class. |
| 4 | How reasoning works; chain-of-thought | Assignment 1 goes out this week — exact dates on Blackboard. |
| 5 | When it goes wrong: hallucination and non-determinism | |
| 6 | Making prompts reliable: self-consistency, structure, citation, self-checking | Assignment 2 goes out this week — exact dates on Blackboard. Midterm problem announced. |
| 7 | System prompts and chatbots; Part One wrap-up | Both assignments wrap up before Fall Break. Fall Break Oct 8–11. Midterm proposal due right after the break — exact date on Blackboard. |
| PART TWO — USING IT ON YOUR OWN PROBLEM | ||
| 8 | Proposals presented; problems assigned to teams | Approved over Fall Break. Presented in class this week. |
| 9 | Workshop: breaking a real problem into steps | |
| 10 | Workshop: iterating until the answer is right | Last day to withdraw (Mon Oct 26). |
| 11 | Workshop: making your solution hold up | |
| 12 | Workshop: building the thing — page, tool, or study | |
| 13 | Workshop: testing whether it actually works | Draft check-in with your section instructor. |
| 14 | Open studio — work on your project with your instructor | Remote instruction day (Mon Nov 23 / Tue Nov 24). Thanksgiving Nov 25–29. |
| 15 | Final presentations | Presentations in class. Report due Fri Dec 4. |
Chapter Two What you do
What this course is
This first-year seminar teaches you to get useful work out of AI systems by writing better instructions: describing a goal clearly, giving the system the context it needs, checking what comes back, and fixing the instruction when the answer is wrong. No programming background is required — writing, argument, and clear thinking are the technical skills here.
What AI actually is, how a prompt is built, the main prompting patterns, how models reason, and the ways they fail. Most sessions are taught from slides. One session each week is hands-on — the lab is done live in class, on the lab page of this site: write the prompt, run it, submit right there. Bring a laptop; if you do not have one, tell your instructor in week one.
You pick a real problem you actually run into, get it approved, and then we solve problems like it together in class, week after week. This half is a workshop, not a lecture. Your final project grows out of the problem you picked.
Labs
Labs run live in the hands-on session each week, on the lab page of this site — your work saves here as you go, and nothing goes to Blackboard. They are graded on completion, not polish — do the work and the points are yours. They are practice, not a test, and they are where the techniques from that week actually get used.
| Rule | What it means |
|---|---|
| Graded on | Completion, in order. Nobody grades your prose. |
| Taken in order | You finish one before the next opens — a lab left undone blocks the next. |
| Due | Done live in the session, on the lab page — submitted there, checked instantly. |
| Late | No late window and no extensions — but finishing late always beats not finishing. |
| AI use | Level 3. Submit your prompts and working results. |
The labs, their steps and their live checks all live on the Labs page.
Grading
| Component | What it is | Weight |
|---|---|---|
| Participation | Being in the room and in the conversation. | 5% |
| Labs | Six labs, completion-graded, taken in order. | 10% |
| Assignments | Two take-home sets, weighted equally — both finished before Fall Break. | 20% |
| Midterm proposal | One page + a three-minute talk on a real problem of yours. | 25% |
| Final project | Build it, write it up (4–6 pages), present it in Week 15. | 40% |
| No final exam. No quizzes. | 100% |
Grade scale
Chapter Three The rules
Deadlines and communication
Assignments are due on Blackboard by the posted deadline. Late work is accepted for three calendar days at 10% per day, then no credit. You get two no-questions-asked 48-hour extensions per semester, requested by email before the deadline — these do not cover the midterm proposal or the final project. Beyond that, extensions are for documented illness, religious observance, University travel, or SAS accommodation. Ask early.
Labs are done live in the hands-on session, on the lab page itself. The three-day late window and the 48-hour extensions do not apply to labs; they are completion-graded, so finishing late is always better than not finishing.
Assignment 1 goes out around Week 3 and Assignment 2 around Week 5; both are due before Fall Break (Oct 8–11), with the exact deadlines posted on Blackboard when each goes out. The design is deliberate: after the break, your attention belongs to the midterm proposal and Part Two — not to leftover assignments.
Academic integrity and AI use
All work is governed by the William & Mary Honor Code. This course is unusual: AI use is the subject, not a shortcut around it. Using AI is expected. What is graded is your judgment — how you direct the system, how you check what it gives back, and whether you can tell the difference. Every assignment is labelled with a level. If no level is stated, assume Level 2.
| Level | What it means |
|---|---|
| 1 — AI-free | Your unaided work. Do not use an AI system to draft, outline, or write it — editing, like grammar fixes, is fine. Rare, and always labelled. |
| 2 — AI-assisted, disclosed | Use AI freely, disclose that you did, and make sure the analysis and the conclusions are yours. Most weekly work. |
| 3 — AI-central | The output is the object of study. Submit your prompts and intermediate results alongside the result itself. Labs and the final project are Level 3. |
You are accountable for what you submit. These systems produce fluent text that is confidently wrong and invent citations that do not exist. Checking facts, quotes, sources and code is your job. “The AI said so” is not a defence, and a fabricated source is a fabricated source wherever it came from.
Never allowed: passing AI work off as unaided where disclosure is required; using AI on a Level 1 assignment; submitting fabricated data, results, or transcripts of things that did not happen; using AI to do someone else’s work or to get around an accommodation.
Do not put personal information — yours or anyone else’s, including classmates’ work or anything covered by FERPA — into a third-party AI tool. Assume anything you type may be kept. If an assignment involves sensitive material, ask for an alternative.
Access and accommodations
Everything required has a free tier. If cost or accessibility is ever a barrier to doing the work, tell your section instructor and you will get an equivalent alternative. It will not affect your grade.
Contact Student Accessibility Services for official accommodations. Your instructors will work confidentially to support approved needs — please get in touch early.
Questions the syllabus doesn’t answer: ask your section instructor. A question before you submit is always good faith.