Generative AI Use Policy

Generative AI Use Policy

Generative AI Use Policy

Summary:

ECE department welcomes, even encourages, responsible and educational use of Generative AI (G-AI) tools for learning new content, improving your understanding, literature search, idea brainstorming, code debugging, etc., so long as the AI generated content is not submitted as graded work as described in this document. All students (and, in fact, all engineers) are ultimately responsible for describing, justifying and/or explaining their work and design, which will be the primary mechanism to determine that any work submitted is, in fact, that of the student.

This document describes the ECE Department policy for appropriate and inappropriate use cases, the proper mechanisms – and examples – for citing AI-generated work, as well as the procedures ECE faculty will use in case of suspected unauthorized use of AI-generated content.

Background and Purpose of the Policy:

Generative AI (G-AI) tools have already transformed—and continue to revolutionize—the way we seek, find, and use information. When used responsibly, G-AI tools can be extremely helpful in retrieving information, solving routine problems, and learning new information. G-AI tools can also be very effective in improving the clarity of writing or summarizing long documents. The ability to use G-AI tools properly and effectively can dramatically improve your efficiency. On the other hand, improper use of G-AI tools can lead to inaccurate information as well as limit your ability to understand and apply important content. 

Our primary goal is to prepare you not just for today’s but also the future’s job market and workforce, where the need for appropriate use of these tools will be all too common, but so will the need for human experts who can verify the accuracy of the G-AI output through the filter of their thorough understanding, analysis and synthesis of the underlying concepts. Furthermore, since a much larger portion of ECE graduates work in classified, confidential, or content-controlled areas, where the use of AI is strictly restricted, it is all the more important to understand how and when to use AI tools properly.

As a result, the ECE Department has devised the following Generative AI Use Policy listing appropriate and inappropriate use cases, the proper mechanisms for citing any AI-generated work, as well as the procedures ECE faculty will use in case of suspected unauthorized use of AI-generated content. Unless explicitly stated otherwise in a course syllabus, the following applies uniformly to all undergraduate and graduate-level ECE courses.

Appropriate and acceptable use of G-AI tools includes the following:

  • Research or learn something new, and/or help your own learning and understanding of any topic.
  • Idea brainstorming for projects, proposals, discussion points, etc.
  • Literature searches to suggest papers or summarize academic content; however, literature review and final citations must always be your synthesis of the reviewed content, and be confirmed from verifiable and reputable sources (e.g., Google Scholar, IEEE, arXiv).
  • For starting or creating a general/routine structure for coding, and/or debugging your code, so long as the substantial/vast majority (75% or higher, not counting the documentation part) of the code is yours, and that *you* understand, test, and document the code entirely and thoroughly. You will be asked to explain your code and demonstrate your understanding, and thereby prove that it is your own work.
  • Routine proof-reading for correcting grammar, typos, clarifying or correcting sentences, so long as the original sentences are yours.
  • Generating non-technical figures to be used as concept illustrations (subject to course-specific permissions - check with the course instructor)
  • Any assignment or exercise where the instructor specifically and explicitly asks you to use an AI tool for a specific task or purpose that is aligned with course objectives and/or outcomes.

Use of G-AI tools for any of the following is inappropriate and not allowed:

The following is a set of examples and not an exhaustive list. Some overlap and repetition among the following items is deliberate and intended for reinforcement. If you are unsure whether a particular G-AI use case is appropriate, please reach out to your course professor.

  • Any use of a G-AI tool that is not properly and thoroughly cited (see below for guidelines on how to cite AI-generated content).
  • Solving or answering any questions in any graded assignment, quiz, exam, or report.
  • Analyzing data, drawing conclusions based on experiments or data analysis to be used in any graded assignment.
  • Answering any essay-type questions (other than for language clarification as indicated above) in any graded assignment, quiz, exam, or report.
  • Determining impact factors, ethical and professional considerations for any design-related course (including Embedded Systems, Engineering Clinics, Professionalism & Consulting, Frontiers in Engineering) that asks you to articulate your design decisions, professional judgment, and ethical discussions in consideration of impact factors. 
  • Code or code fragments that exceed 25% of the total script for any programming assignment or any project that requires you to develop code. If the instructor provides you with a base code to modify or to add to, then the modifications or additions cannot be AI-generated and must all be your work.
  • Uploading class/course material, lecture notes to any unauthorized G-AI tool (e.g., ChatGPT and others), or online learning platform (including Chegg, CourseHero, and other similar platforms).
  • Mathematical derivations submitted as part of a graded assignment, quiz, exam, or report.
  • Algorithm design submitted as part of a graded assignment, quiz, exam, or report.
  • Using AI to generate any part or section of a lab report, project report, or any other submission. For example, uploading raw data, results, or incomplete notes into AI to generate a report is not permitted.
  • Using AI to interpret, process, or analyze results from simulations.

It is worth noting that:

  1. Pretending or presenting any content as if it is yours that is, in fact, generated by others (including Gen AI) is plagiarism. Paraphrasing or rewording the borrowed content to hide its original source is not only plagiarism, but also academic dishonesty. Either case is a form of academic integrity violation that will not be tolerated and will be reported to the Provost’s Office in accordance with Rowan University's Academic Integrity Policy.
  2. Lecture notes made available to you are the intellectual property of the course developer/instructor. These lecture notes and materials are copyrighted and protected by copyright law, regardless of whether a copyright notice is visible (as is the case for any work of authorship under US law). When provided to you (such as on Canvas), you are merely given the very limited right to use them for your own personal learning. Therefore, uploading any course content to any online portal, including online learning platforms such as Course Hero, Chegg, etc., or any Gen AI platforms (such as ChatGPT, CoPilot, Gemini, etc.) is illegal and strictly prohibited. If detected, it will be reported as an Academic Integrity Violation at Level 3 or Level 4. The instructor (or content owner) may also independently take legal action against you under Copyright Law.

    Simply stated: do not upload any course materials/content to any unauthorized site without the express written permission of the instructor/copyright holder.

Citing G-AI tools:

In all cases, the specific use of a G-AI tool must be properly and completely cited as follows:

  • Citing routine use of a built-in G-AI tool to correct individual errors for grammar, typos, clarification, etc., use the following format with only one citation for the entire document.

<Company name)>, <Name of the tool>. Accessed on <date in which you
         used the tool>, Model or Version: <model or version used>, <descriptor>. Available at:
       <link to the main page for the tool>. Context: provide a brief description of the use.

  • Example 1: Grammarly Inc, Grammarly. Accessed on 15 June 2025, Version: 14.1232.0 (browser), large language model-based writing assistant integrated into Word. Available at:  https://www.grammarly.com. Context: used for routine grammar /typo correction.
    • Example 2: Google, Gemini. Accessed on 20 July 2025, Model: Gemini 2.0 Flash, large language model-based writing assistant integrated into Google Docs. Available at: https://gemini.google.com/app. Context: used for routine correction for grammar and typos, and obtaining a more concise writing.
  • For all other and more substantial uses of G-AI tools, use the following format, and include a separate citation for each use. If you are using up to three prompts, include all. For longer conversations, include the first, second, and last prompt with ellipses (...) in between.  For additional clarification, when needed, provide context. If asked by the faculty, include the entire transcript as an appendix.

<Company name as author)>, <Name of the tool>, “<exact prompt used as title.>”
Accessed on <date in which you used the tool>, Model or Version <model or version used>, <descriptor>. Available at: <link to the main page for the tool>.  (if applicable, Context: provide a brief explanation), (If applicable, add: complete transcript provided in Appendix).

  • Example 1: OpenAI, ChatGPT, “Can you explain the practical meaning of eigenvectors and eigenvalues with examples at a level appropriate for someone who is in the second year of an engineering program?,” Accessed on 1 July 2025. Model GPT-4o, large language model-based AI chatbot. Available at: https://chatgpt.com/
    • Example 2: Microsoft, CoPilot, “Help me understand Maxwell’s equations”, “Can you explain each equation with more context?”, …, “Which of these equations directly apply to an induction stove, and what is the principle of operation?” Accessed on 10 June 2025. Version Copilot Pro, large language model-based AI chatbot. Available at: https://copilot.microsoft.com/
    • Example 3: Google, Gemini, “When plotting my FFT in Matlab, the horizontal axis simply shows the index. How do I get it to show the actual frequency?” Accessed on 17 June 2025, Model Gemini 2.5 Pro (preview), large language model-based AI chatbot. Available at: https://gemini.google.com/app.  Context: partial code-fragment provided by this tool is used in FIR filter design, complete transcript is available in Appendix A, the lines obtained from this tool are indicated in the code.
    • Example 4: OpenAI, ChatGPT, “Can you provide base code in C++ for the Arduino IDE to run on an ESP32 Nano that monitors two digital inputs and sets two digital outputs based on their states? I’m a second-year ECE student learning embedded programming,” Accessed on 13 June 2025. Model GPT-4o, large language model-based AI chatbot. Available at: https://chatgpt.com/. Context: partial code-fragment is used in creating a base code for Arduino, complete transcript is available in Appendix B, and the lines obtained from this tool are indicated in the code.

What will happen if the ECE faculty suspects inappropriate use of AI-generated content?

  • Students are always responsible for explaining, justifying, demonstrating their work, in particular for any design, coding, analysis, or synthesis work submitted for grading. If academic dishonesty is suspected, the faculty will provide an opportunity for the student to justify, explain, or demonstrate their work or design in a subsequent in-person oral or written format. If the student cannot adequately explain, describe, or otherwise demonstrate their understanding of the submitted work, this incident will be considered an intentional academic integrity violation and will be reported as a Level 3 or Level 4 violation.
  • For any assignment that is graded for writing (such as lab or clinic reports), ECE faculty may compare your writing style obtained in an in-person writing assignment to the one in the suspected assignment or may ask you to (re)write a paragraph or two, in person, on the same topic to compare the writing styles. The faculty may also ask you to explain, clarify, demonstrate that the writing is yours. If, based on this comparison, the ECE faculty believes that the work submitted for grading contains more AI-generated content than is allowed, it will be reported as an academic integrity violation, at a level appropriate to the amount of plagiarism detected.
  • Not citing or improperly citing the use of AI-generated content will be considered cheating/plagiarism and will be reported as an academic integrity violation at a level appropriate to the nature and amount of the plagiarized content.
  • Note that fake citations, references that do not actually exist or are not relevant to the assignment, or other forms of hallucinations will be considered as evidence of AI-generated content (or at the very least, content not genuinely generated by the student submitting the work). Such cases will be reported as an academic integrity violation at a level appropriate to the nature and amount of the plagiarized content.

References:

[1] Grammarly Inc, Grammarly. Accessed on 13 June 2025, Version: 14.1232.0 (browser), large language model-based writing assistant integrated into Google Docs. Available at:  https://www.grammarly.com. Context: used for routine correction for grammar and typos.

Appendix A - Transcript for the FFT plotting chat with Gemini

User: When plotting my FFT in Matlab, the horizontal axis simply shows the index. How do I get it to show the actual frequency?

Gemini: When plotting an FFT in Matlab, the horizontal axis (x-axis) will indeed default to showing the sample index. To display the actual frequency values, you need to calculate the corresponding frequency vector. Here's a breakdown of how to do it, along with explanations and common considerations:

The Discrete Fourier Transform (DFT), which the FFT algorithm computes, transforms a signal from the time domain to the frequency domain.

  • Y = fft(y): If y is your time-domain signal, fft(y) returns a complex vector Y representing the frequency components.
  • Sampling Frequency (Fs​): This is the number of samples taken per second. It's crucial for mapping indices to actual frequencies.
    …
    f = (-N/2 : N/2-1) * (Fs/N); % Frequency range from -Fs/2 to Fs/2
    Y_shifted = fftshift(Y); % Shift zero-frequency component to the center
    …

Appendix B - Transcript for ESP32 Nano Base Code for Arduino chat with ChatGPT

User: Can you provide base code in C++ for the Arduino IDE to run on an ESP32 Nano that monitors two digital inputs and sets two digital outputs based on their states? I’m a second-year ECE student learning embedded programming.

ChatGPT: Sure! Here's a minimal base code example using digitalRead() and digitalWrite() functions. It monitors GPIO 4 and GPIO 5 as inputs, and controls GPIO 16 and GPIO 17 as outputs based on those inputs …
// Define input and output pin numbers
const int inputPin1 = 4;    // Digital Input 1
const int inputPin2 = 5;    // Digital Input 2
…

User: Thanks! Can you clarify what happens if the inputs float? Should I enable internal pull-ups, or do I need external resistors?
ChatGPT: Great question. For stable readings, you should either connect external pull-down resistors or enable the ESP32’s internal pull-up resistors using pinMode(pin, INPUT_PULLUP). I’ll modify the base code to reflect that...