Using ChatGPT Effectively and Ethically

There are several ways, such as using AI to answer questions about an assignment prompt or asking for examples of how ChatGPT might answer questions. For instance, if an assignment prompt doesn’t make sense, you can copy the prompt into ChatGPT and ask for an explanation. You might also be able to gain a better understanding of an assignment by asking for the bot to create several examples of how the assignment might be completed. Remember, ChatGPT will not always provide the correct answers, but it should provide a reliable format of the genre you are writing in.  

The classrooms, workplaces, and organizations of which you are a part are communities that have their own best practices and ways of being and behaving. In some classes, instructors may prohibit the use of generative AI tools in some cases; alternatively, an organization you intern for may be avid users of those same tools. Knowing the norms and best practices of whatever community you’re a part of will help you avoid mistakes. 

But ethical use isn’t always about avoiding mistakes – it’s often about intent and a lack of transparency. What is your intent in using ChatGPT? Is it to help you learn or to avoid learning? If you’re using it at work, is it to do a task more efficiently so you can focus on deeper, more productive endeavors that add value to the firm or are you using it to do the bare minimum as quickly as possible? If you use ChatGPT for class, are you citing it as a source? If you are using it at work, are you making sure that the information it provides is credible? Are you supplementing ChatGPT as necessary or are you blindly accepting its outputs?  

Ethical use of generative AI tools is still very much in flux, which is not surprising given how new these tools are. We will update this resource regularly as best practices emerge, but focusing on your intent and being transparent are useful starting points. 

AI language models generate responses based on patterns of information from the data they have been fed but AI doesn’t always know where this information came from, and the bot has a tendency to invent citations. This means that students using ChatGPT need to investigate where ChatGPT has gotten its information and check to make sure the source is reliable.  Changing a few words of the generated text does not make the text unique any more than if you changed a few lines in a text someone else wrote. If you use ChatGPT and want to avoid plagiarism, the simplest thing to do is cite your usage properly. 

If you’re using AI as a search engine, an outlining tool, or to review your grammar and usage, then no citations will likely be necessary unless it significantly rewrites your text.  

For tips on avoiding plagiarism in general, please consult resources provided by Student Success and by the Business Communication Lab. 

As generative AI is a relatively new development, the university is still determining the best solution and policy response for students and faculty. That said, the university’s plagiarism policy and sanction rubric clearly state that students must cite correctly (see Level One Violations) and must have the instructor’s authorization to use a resource before using it (Level Zero Violation). As before, we will update this resource regularly as university policies solidify and/or change. Faculty members who think a student has used generative AI without citing or after they specifically request students not use it can use the university’s institutional-grade resources for detecting AI-generated text. Please contact the Office of Academic Initiatives and Integrity directly regarding next steps and access to these tools. 

It depends on the question. ChatGPT, for example, does well with straightforward and factual questions; examples abound from the legal world and medical world regarding its veracity. The platform, however, is not connected to the internet and thus has limited knowledge of the world and events after 2021.  

If your questions involve recent events, ChatGPT may provide you with inaccurate information. These inaccurate responses, known as hallucinations, occur because ChatGPT generates text that conforms to what its training models indicate an answer or response to your questions and prompts should sound like. ChatGPT was trained on an enormous amount of information on the Internet, and while the Internet is full of credible information, it has a lot of garbage: hate speech, untruths, biased information, and deliberate misinformation.  This garbage found its way into the training models and thus can sometimes affect the responses given to users. As such, you should always check the sources generative AI provides and make sure they are actual, credible sources.  

You should start experimenting and learning how these tools work because you’ll likely see them and perhaps use them in internships and as employees after graduation. Article series about generative AI platforms as well as tutorials abound, as do tips on searching for jobs and on replacing mundane tasks. 

That said, you should avoid providing private or confidential information to ChatGPT or other generative AI tools. These tools log every conversation you have with the platform and then share these conversations with other companies and its own AI trainers. So, if you type any personal or sensitive information into the dialog box, know that per its current privacy policy, that information it will be recorded and saved on its servers.