Ethical Guidelines for Creating Content in a World of Generative AI: For the Classroom and Beyond
I was preparing the syllabus for a social computing course I’m teaching next fall when I saw Cori Faklaris’ blog post about her course policy for using AI tools. It helped inspire me to write down ethical guidelines that I’ll be providing my students on: 1) using generative AI tools, 2) maintaining integrity and honesty in their work, and 3) acknowledging other people’s work.
While these guidelines are aimed at college students, they may be helpful to anyone thinking about using AI tools in their work. (Although the last two sections are more focused on traditional writing practices.)
Research has shown that the tone a syllabus takes affects students’ perceptions of the instructor. Syllabi that take a more friendly tone (versus unfriendly), lead students to perceive the instructor as being more approachable and motivated in wanting to teach the course. I also don’t want this language to be perceived as a punitive policy (though in extreme cases, it may be), but rather as a moral compass that will help students differentiate between ethical and unethical creative practices. The goal of these guidelines, then, is to foster an ongoing discussion between my students and I. I fully expect some of these guidelines to change over time.
Using AI and Other Creative Tools
In a world increasingly saturated by AI tools, I want to give my students a strong technical and ethical foundation for using them. I believe that they will and perhaps should use them, though there may also be times where such tools are detrimental to their learning. I hope these guidelines below will help explain how to use AI tools in an ethical manner, while also helping them learn the pitfalls of over-reliance and misplaced trust.
If you use AI or other creative tools (including but not limited to Microsoft 365 Copilot, Microsoft Bing Chat, GitHub Copilot, Open AI’s ChatGPT, Facebook’s LLaMA, Google’s PaLM, Grammarly, and Canva) to generate, draft, create, or compose any portion of your work, in a footnote or separate section please:
Identify what part of the work is from yourself and what part is from the AI tool, and acknowledge the human labor involved in developing AI models and tools. If you use AI-generated content or ideas, please don’t claim credit for them. AI models are trained on the work of thousands — if not millions — of under- or unpaid writers and artists, and further refined by crowd/gig workers. By crediting the AI tool and acknowledging the underlying human labor you are, at least to some extent, giving credit where it is due, just as you would want someone to credit you if they used your work later on. An acknowledgement is hopefully a small step towards addressing its extractive practices.
Explain how the AI tool’s output helped improve the quality and/or creativity of your work. Remember that AI models can only reproduce the past, and can’t imagine new possibilities — whereas humans can. Instead, leverage the outputs of AI tools as a starting point for your creative process, not an end point. It may even be helpful to spend time brainstorming and creating on your own before allowing your ideas to be influenced by the outputs of AI tools. Note that by overly relying on AI tools, you may also be limiting your future self from developing the skills (and neural pathways) needed to be creative on your own: generating new ideas, synthesizing disparate ideas, writing thoughtful prose, and designing other types of content. You don’t want AI tools to become crutches*.
Fact check any claims you included from AI tools, and cite any primary/original sources. Remember that AI-generated content is frequently misleading or factually incorrect due to inherent limitations in how these models are designed and currently work.
Note:
1. Also see the MLA’s recently-released guidelines on formatting citations to content generated by AI tools.
2. The first two points borrow heavily from Cori Faklaris’ blog post, but the descriptions are my own, as is the nod to acknowledging the underlying human labor.
3. * Though this is a useful analogy meant to warn against overreliance, I want to acknowledge that it is an ableist term.
Maintaining Integrity and Honesty in Your Work
Below is a tentative set of guidelines to help students think about how to maintain the integrity and honesty of their work:
In all cases, some of these actions may cause others to lose trust in you and reduce the public's now-fragile trust in the scientific process. Others who use your work or underlying data may also be harmed: by having their time, money, and efforts wasted. You may also end up causing reputational harm to yourself, your institution, and others. Finally, you may be limiting your future self by not doing the work needed to develop your skills.
Plagiarism
Submit your original work. More precisely, don’t claim someone else's work, language, ideas, or arguments as your own if you know it's not. The goal of any course is to help you grow personally and professionally, but you submit someone else’s work, you’re hurting yourself. It's also not fair to the original author for you to get credit for their work.
Self-plagiarism
Submit work that you created for this course. Don’t submit work you’ve already been credited for — you’ll be affecting your own personal and professional growth. It's also not fair to others to get credit twice for the same work, unless you explicitly cite your prior work.
Impersonation
Be yourself. Never pretend to be someone you’re not; although pseudonyms and pen names may be acceptable in certain contexts. Don’t submit your work on others’ behalf. It's not fair to your classmates, the person you may be trying to help, or yourself.
Forgery
Follow the law. Don’t produce or submit a legal document (e.g., a doctor's note, a signature, a check, etc.) unless you’re legally authorized to do so. Doing so may cause physical, emotional, financial, and reputational harm to others and yourself.
Falsifying Data
Make sure you use data appropriately. Don’t make up, alter, or incorrectly reuse data that you use in your work. This includes — partially or entirely — using made up or fictitious data, modified data that alters the underlying characteristics of the data, or reusing data from a prior study without proper justification and citation.
Misrepresenting Findings
Make sure you interpret your findings correctly. Don’t willfully misrepresent or mischaracterize findings that you arrived at by analyzing the data. Double check your analyses and writing to make sure you are accurately describing what you found (or did not find). In addition, don’t: a) omit or include data without clear justification or b) search for plausible results (i.e., p-hacking) when there is no underlying pattern.
Acknowledging Other People’s Work
To help maintain your own integrity and honesty, remember to follow proper citational practices, that is, referring back to others’ work in your own.
Having Original Ideas
You will come up with an idea on your own that, perhaps, someone else came up with at the same time or over a century ago. It happens all the time — that’s okay! If you do come up with a novel idea, you may still need to cite others’ work that helped inspire your idea. Oftentimes you may not even remember if any idea was entirely your own or partially came from someone else. This is a classroom setting, so it’s okay if you forget to cite some sources.
If you later realize your idea is the same as or similar to someone else's work, you should consider citing their work. If you don't, readers will still probably notice it and take you less seriously. If you take all the credit for an idea, it’s not fair to the person who came up with the idea first.
Citing Others’ Work
When you include parts of someone else's work (an idea, an argument, a quote, etc.) in your own, remember to provide a reference to their work. This way, others receive credit for their work, and you (are hopefully able to) provide additional justification for yours.
Citations (and papers) have politics. When citing other people’s work, you may want to think about citational justice — in other words: Who are you citing? Are you overlooking any relevant and important prior work? Are you citing research(ers) from only one particular field, university, or country? Are you citing researchers of a certain gender, race, or ethnicity disproportionately over another (within reason and within the context of your work)? Are your citations cursory (i.e., a simple nod to prior work) or are you deeply engaging with that prior work?
It’s also important not to mis-cite others work, that is, using someone else’s work to make an argument or provide evidence when that work actually provides an opposite or orthogonal (unrelated) argument or evidence. You want others to take your writing seriously, so you should take others’ writing seriously as well. Who knows, maybe the author of an article you cite ends up reading that very sentence — it happens more often than you think.
Quoting and Paraphrasing
If you use someone else's words in your work, remember to either: 1) paraphrase and cite OR 2) quote and cite their work. Generally use direct quotes if you have a good reason to do so: because it makes a certain point better than you could; because the words are so moving or make a historical point; to show authority; or to critique or comment on those words. Otherwise, paraphrasing is probably more appropriate. Like mis-citing, it’s important not to misquote someone else’s words.
If it isn’t academic or professional writing, it may be okay to use someone else’s words without quotes (for better readability), but still remember to cite them.
Feedback
Like I said at the beginning, these guidelines are meant to start a discussion and not be an ultimatum. I welcome any and all feedback you might have — leave a comment down below!
Adapting These Guidelines
This work is licensed under a Creative Commons License (CC BY 4.0). You can copy and redistribute the material in any medium or format, as well as remix, transform, and build upon the material for any purpose.
To incorporate these points into your own syllabus, please: (1) cite this article, and (2) send me an email (sukritv+technomoral@uw.edu) — it’s always great to hear how people are using my work :)
To cite this article:
Sukrit Venkatagiri. Ethical Guidelines for Creating Content in a World of Generative AI: For the Classroom and Beyond. Technomoral, March 18, 2023, https://technomoral.substack.com/p/ethical-guidelines-for-creating-content
References and Updates
Some of these guidelines draw heavily on my conversations with other people, other people’s writing, as well as my own thoughts and ideas over the last few years.
The first section, “Using AI and Other Creative Tools” draws heavily on Cori Faklaris’ blogpost, Kristopher Purzycki’s blogpost, and my conversations with Tim Gorichanaz. However, my guidelines take a less punitive tone and provides context as to why I am requesting my students provide additional detail.
The sections on “Maintaining Integrity and Honesty in Your Work” and “Acknowledging Other People’s Work” were written entirely by me, though may implicitly draw upon my prior conversations, readings, and ideas. The third section additionally relies on Neha Kumar and Naveena Karusala’s term of citational justice in HCI that helped solidify in writing my own perspective towards fair citational practices.
3/23/23: Updated a reference to include MLA’s new guidelines published on March 17th for citing generative AI content.
This is a valuable addition to my own library of perspectives on using generative AI in academia. I may also be leaning on your work here for my own teaching (with proper attribution of course)!