Welcome to Cinema Perspectives on Artificial Intelligence!
Like most 290/390/590 courses, this is a new/experimental course, and there are still a lot of details being figured out! This site is up early to give students a place to find more information, and will be updated over the next few months.
In this course, we will use cinema as a lens on society’s view of artificial intelligence. We will discuss how this view has changed over time, and how it compares to the real-world status of AI. It will also include a unit on the usage of AI in creating cinema, and its accelerating impact on moviemaking.
Students should be prepared to watch one movie a week, and are expected to participate in class discussions. There will be four projects over the semester.
The projects are the component most in flux as of the time of writing this. However, we plan to have at least one in each of these categories:
- Paper
- Optional coding (for those who have programming beyond 201 and are interested)
- Optional film-making
- Interacting with AI
The very rough breakdown of the course is first looking at the historical perspectives on AI since one was first portrayed in cinema almost a century ago, then an intermission on how AI is being used in filmmaking, then looking at more modern (last 20-25 years) representations of AI. Throughout the course, we will be learning some of the real-life history of AIs and how a selection of modern systems work.
We will exploring one or more AIs that generate videos, such as Sora, Runway, and Luma. The landscape of what is available to the public and how much they cost to use is constantly changing, so we will be deciding which ones to use closer to when we need them.
Details
TueThu 3:05-4:20PM in LSRC A247
Optional synchronous film screening — time and room TBD
First class meeting: Thursday, January 9, 2024
Instructor: Kate O’Hanlon
Grad TAs: Jiajun Chang, Shariar Vaez-Ghaemi
UTAs: Ondine Peck-Voll, Allie Brown, Camden Chin, Carlos Penzini, Nico Gonzalez
Office hours:
TBA
Description
See above.
Prerequisites
The only prerequisite is CompSci 101 or equivalent. This can look like EGR 103, AP courses, being self-taught, or a variety of other coding backgrounds. The programming language is not important, but should be more advanced than block programming or Alice.
Those with a stronger background in CS/AI or in cinema are very welcome. A large portion of class time will be spent discussing and analyzing the films together, so having a variety of perspectives and expertise levels in the room is excellent!
Book/Readings
There will be required readings for this course. We will do our best to make them all available to students for free, but are still working on the logistics with Duke Library and other resources.
Grading:
Participation (in-class participation, surveys, weekly responses): 50%
Projects: 50%
Schedule
A more detailed schedule will be posted here soon.
This schedule is a plan, not a commitment. Depending on the time needed to cover the various topics, it may be necessary to skip, rearrange, or substitute some of the material listed.
Week 0 (January 9) has no required viewings or readings in advance of the first class meeting.
Week 1 will start off with a recent film to begin our discussions of AI usage in creating cinema and some of the questions around it.
Weeks 2-7 will walk through historical representations of AI in cinema. At the same time, we will focus on some core questions: What is AI? What does it look like? Should we trust it?
Weeks 8-9 will focus on the usage of AI in creating cinema.
Weeks 10-14 will look at modern mainstream cinematic interpretations of AI. At the same time, we’ll look at how many modern tools work and discuss whether modern film depictions are at all accurate or likely in the near future.
