Table of Contents
Unlike most content in days past, generative artificial intelligence (AI) content is created using text materials, images, sounds, graphics, and other inputs from millions if not billions of sources. This breadth of information makes knowing how to cite AI tools and sources problematic. However, the need to give credit where credit is due has not changed. If you didn’t write it, you need to cite it.
Why Citing AI Tools Matters
Citation matters because it is fair, honest, respectful, and responsible. Those factors sum up academic integrity. Doing otherwise would be considered academic dishonesty, which could lead to negative consequences. In the case of generative AI tools, whether they create text, images, or music, citation gives credit to the creators of the tool, the algorithms it makes use of, and the datasets it is trained on. It recognizes both the technology and the underlying resources.
Citation is proper ethical practice, and it fosters a culture of considerate AI usage. When citing AI tools, you communicate their power by showing what was used and how. You also help others in their AI-assisted endeavors. Citing AI tools matters because it promotes an informed and collaborative environment that leads to shared benefit.
Types of AI-Generated Content We Might Need to Cite
When it comes to citation, you might ask yourself if it matters how you made use of the generative AI tools. Is there a differentiation between using the tools for creation of content versus using the tools for analysis, support, or understanding? Actually, yes!
When you use AI to create content, you use it as a material development tool. Whether you paraphrase, quote, or incorporate content (whether text, image, data, or other) created by the tool into your own work, it should be cited.
When you use AI for analysis and understanding, you are not creating new content but are instead using it to understand existing topics and materials. Using AI as a learning tool does not need citation. In some cases, though, you still may want to acknowledge any functional uses of the tool, like editing, translating, or giving design ideas.
Remember to Check with Your Instructor
Always confirm with your instructor whether AI tools like ChatGPT are allowed for each assignment or for assignment preparation. It is possible your instructor could view unauthorized use of AI tools as an academic integrity offense equivalent to plagiarism, misrepresentation, fabrication, or similar.Check with Your Instructor
Citing AI Tools
The use of generative AI tools almost always necessitates some type of citation. Your citation may include identifying the AI tool’s name and developers, specifying the version and parameters used, mentioning the date and time of usage, and describing how the AI tool was used. How you cite AI tools will vary based on the material.
Written Content
Content generated from AI tools, like ChatGPT or Dall-E, can be impossible to recover due to the AI’s generative process that functions via computer algorithms. While some sources state you should refer to it as a personal communication, it is often preferable to use in-text citation with a coinciding reference list entry.
Citation styles are as follows:
American Psychological Association (APA)
If you directly quote material, put quote marks around it, followed by the author, page number if applicable, and year of publication in parenthesis. If you are not directly quoting the material but are using it in your own synthesis of thoughts, follow the paragraph with the author and year of publication in parenthesis.
- In-text citation format: Content (author, year).
- In-text citation example: According to ChatGPT, the sky is blue (OpenAI, 2023).
- Reference list format: Author. (year). Tool name [Large language model]. URL
- Reference list example: OpenAI. (2023). ChatGPT (version 3.5) [Large language model]. https://chat.openai.com/
Modern Language Association (MLA)
Any material written by AI, whether directly quoted or not directly quoted, should have a description of the AI prompt used in parenthesis after it.
- In-text citation format: Content (“Description of AI prompt used”).
- In-text citation example: According to ChatGPT, the sky is blue (“What color is the sky”).
- Note: If the prompt used is too long for an in-text citation, add it as a parenthetical citation or a footnote.
- Works cited format: Description of AI prompt used. AI tool name, version, company that made the tool, date, URL.
- Works cited example: “What color is the sky” prompt. ChatGPT, version 3.5, OpenAI, 17 Aug. 2023, chat.openai.com.
Artistic Visual Works (images, graphics, design)
Like with written content, visual works from AI tools like Craiyon can be impossible to recover due to the generative nature of AI. To cite AI tools for media, include a caption that specifies or summarizes what you used to create it and when.
American Psychological Association (APA)
Create a caption under the image or figure.
- Caption format: AI tool name. (Year). Company that made the tool (version) [Text-to-image model]. URL.
- Caption example:
Fig. 1. “A blue sky”
Craiyon. (2023). Craiyon LLC (Aug 16 2023 version) [Text-to-image model]. https://www.craiyon.com/
Modern Language Association (MLA)
Create a caption under the image or figure.
- Caption format: Description of AI prompt used. AI tool name, version, company that made the tool, date, URL.
Caption example:
Fig. 1. “A blue sky” prompt. Craiyon, 17 Aug. 2023 version, Craiyon LLC, 17 Aug. 2023, http://www.craiyon.com.
Audio
The citation rules for audio are less clear, but it is generally wise—and ethical—to disclose your use of AI tools. This is particularly true if you did not edit or revise the generated audio content.
Suggested Template: Name of AI Tool. (Year, Month Day you generated the content). Exact text of question or prompt you entered [AI-generated text/image/video, etc.]. Name of Company/Developer if different from name of AI tool.
Example: Beatoven.ai. (2023, August 16). Indie happy 10 second [AI-generated music]. Beatoven Private Limited.
Computer Code
The rules for citing computer code are the least clear, but again, the ethical thing to do is disclose your use of AI tools, even if they were used for limited coding modification. To achieve this, you can include a comment within your code that explains the AI tools you utilized and the specific manner in which you applied them. The comment could be in the header of the file or script, or whenever it is used within your code at a function, variable, or method level.
Code header example:
# Generative AI Attribution
# Model: GPT-4 by OpenAI
# Description: This code utilizes the GPT-3 model to generate [content here].
# Reference: OpenAI. (2023). GPT-4. https://www.openai.com
Code body example:
# This code was generated using ChatGPT version 4 on August 16, 2023.
import OpenAI
input_prompt=”my prompt here”
generated_text = openai.Completion.create(engine=”davinci”, prompt=input_prompt)
print(generated_text.choices[0].text)
You can also add a detailed description in your formal documentation of how you used generative AI tools during code development.
Final Thoughts
Attribution is important because it gives credit where credit is due. Citing AI tools ensures you are being as transparent, responsible, and respectful as possible regarding where your materials came from. When you take the time to disclose your sources, you are choosing to act honestly and with integrity, which is something we should all strive for.
Additional Resources
- How do I cite generative AI in MLA style?
- How to cite ChatGPT
- 8 Ways to Use Generative AI for Student Assessment
About Dr. Anne Arendt-Bunds
Dr. Anne Arendt-Bunds holds a Bachelor of Arts (B.A.) degree in English, a Masters in Business Administration (M.B.A.) from the University of Minnesota Carlson School of Management, a Master of Science (M.S.) degree in Educational Change and Technology Innovation from Walden University, and a Doctorate of Education (Ed.D.) from Utah State University, with an emphasis in higher education.
Additionally, she is a certified Six Sigma Black Belt through the American Society of Quality (ASQ) and was previously certified with Project Management Professional (PMP) certification through the Project Management Institute (now expired). Dr. Arendt-Bunds’ love for learning fuels her aspirations to inspire and support other students in cultivating their own passion for education.
