Notes on LLM use and Academic Honesty

Overview/Summary

  • Use of ChatGPT and other LLMs to generate code, or to generate, summarize or translate comments or other writing is not permitted in this course. Doing so will be considered Academic Dishonesty and the policies outlined below (and in the course syllabus) will be applied. Any code or text included in your assignments or in communications related to this class must be presented in your own words.
  • If code is included in assignments that comes from external sources and is not attributed properly, this will be considered Academic Dishonesty. The policy regarding Academic Honesty is included below. Information about proper attribution can be found on this page.
  • You do not need to provide citations for anything that you have learned in this course via lectures or the course’s OpenProcessing online materials.
  • If you have used the p5.js reference site to learn about new functions, please include the URL and a description of what that function does in your comments.
  • If you are including code that you have found from a source outside of the p5.js reference, you must provide detailed notes on why you are using that code, and how that code works. You should first ask yourself whether the problem can be solved using the skills we have explored in lectures and in the course materials before solving problems using external resources. The assignments are designed so that you can accomplish all of the requirements using only the resources that we have learned in the class. When I am grading, I will take this into account – an assignment that uses the skills we have learned in class to solve problems will always receive a better grade than assignment that uses existing code to solve that problem. This is why I strongly advise against LLM use - it will (maybe) solve the problem for you, but you will learn nothing in the process.

Please make yourselves aware of the following sections of the course syllabus (which is unchanged from the beginning of the semester):

Academic Honesty and Plagiarism

Violations of academic integrity are considered to be acts of academic dishonesty and include (but are not limited to) cheating, plagiarizing, fabrication, denying other access to information or material, and facilitating academic dishonesty, and are subject to the policies and procedures noted in the Student Handbook and within the Course Catalog, including the Student Code of Conduct and the Student Judicial System. Please note that lack of knowledge of citations procedures, for example, is an unacceptable explanation for plagiarism, as is having studied together to produce remarkable similar papers or creative works submitted separately by two students, or recycling work from a previous class.

Please review NYU Tandon’s academic dishonesty policy in its entirety. Procedures may include, but are not limited to: failing the assignment, failing the course, going in front of an academic judicial council and possible suspension from school. Violations will not be tolerated.

All work for this class must be your own and specific to this semester. Any work recycled from another, non-original source will be rejected with serious implications for the student. Plagiarism, knowingly representing the words or ideas of another as one’s own work in any academic exercise, is absolutely unacceptable.

This includes copying code for other sources, using code from other sources with only slight modifications and using code from other sources without a reference. It is very easy to find code examples on the internet, especially for beginner-level p5.js concepts. It is very important that you write your own code for the fundamentals portion of the semester (weeks 1-10) and do not copy and paste code from other sources. This is a matter of fully understanding the fundamentals of coding – if you rely too heavily on copying and pasting code from examples, you will not be able to write code independently and fluently. Every assignment you submit until week 10 should be hand typed by you.

For your advanced assignments and final project, there may be instances where you wish to adapt an example or use a snippet of code found online. You must cite the work and author and comment each line of code to explain what it is going on programmatically. Each comment should describe what the specific line of code is doing. The expectation, when working from examples, is to apply code examples towards unique and novel ideas. If you are using borrowed code in your assignments, its use should not determine the look and feel of your work. It should, instead, be modified to help you achieve your own personal creative vision. If you use code from online sources, I expect to see a link to its source in the comments. Failure to attribute other peoples’ code will result in failure of the assignment. Likewise, you should never borrow more than 30% of your code from other sources, even if properly attributed.

Statement on LLMs / ChatGPT / etc

The primary objective of this class is to equip students with coding skills, so that they can advance in this program and in their future careers as independently-capable programmers. Learning in this class is structured around assignments that challenge students to synthesize course material through creative problem solving and creative exploration. Put simply, you will not learn how to code if you rely on LLMs to do your work for you. And, since the point of this course is to teach you how to code, you will not succeed in this course if you rely on LLMs to complete your work.

Students should not use ChatGPT or other AI assistants to:

  • Generate project content based on assignment prompts
  • Generate code based on assignment prompts
  • Comment or modify existing code to match assignment specifications
  • Translate or “clean up” your code

I read and review all assignment code in detail. There are many telltale signs when work is generated by ChatGPT or LLMs, such as the inclusion of outdated or rarely-used syntax, use of techniques that exceed the scope of this course, and sprawling, inaccurate code. When tasked to respond to creative prompts, LLMs produce generic results that are not reflective of students’ personal artistic voices. Effectively, code generated by ChatGPT is equivalent to student work that is not engaged in learning, nor reflective of creative intention. If your work reads as though it was written by AI, it will receive no more than a C- grade. If you require assistance with coding, please reach out to me. I am happy to meet with students to work through problems, or to point you towards resources that can help you to learn better coding habits.