While generative systems like ChatGPT are not going to write a paper for you (or at least not one you’d want your name attached to), they certainly can be helpful in conducting research and organizing your submissions. Asking a generative system directly for information on a topic is not guaranteed to produce reliable information, but it can be useful in a variety of tasks.
When an assignment allows you to select a particular topic to write about, a generative system may be able to brainstorm a list of potential ideas based on your prompt. Prompts can be made broader or more specific depending on the assignment and your field of interest. Of course, you’re not obligated to use a suggestion exactly as written; they can be altered based on your preferences and on the requirements of the assignment.
EXAMPLE:
If asked, a generative system may be able to provide a list of possible readings and sources relevant to your research topic. This should be used with caution, however: generative systems have limited date ranges and resources to work with, and may not suggest material that is entirely appropriate for your field of study (in the below example, several articles have publication dates that are over 20 years old, and one article was only tangentially related to the given research prompt). Generative systems may even make up sources that don’t exist, so always use your best judgment and personally review any and all sources you are considering using for your assignment(s).
EXAMPLE:
If you already have a research topic and some relevant sources, you could ask a generative system to construct a rough outline of your paper, as a way of organizing your thoughts. Keep in mind that you are free to alter any results to your liking, and that a finished paper need not follow a computer-generated outline exactly.
EXAMPLE:
Generative systems may have some capability at formulating bibliographic citations for your works cited pages. However, it is important to note that they may not be reliable in this regard. If you do choose to generate citations this way, be sure to take the following steps:
-1: Specify exactly what kind of citation you want. Make sure to specify the style (MLA, Chicago, APA, etc.), the format (bibliographic, in-text, footnote, etc.), and any relevant information about the source itself.
-2: Make sure the source falls under the parameters of the generative system itself, for example that it wasn’t published or created outside of the date range of the system’s parameters. For example: ChatGPT-3’s pre-trained model does not extend to information generated later than September 2021, so it will have no access to anything published after that time.
-3: Independently verify that the system generated a correctly formatted citation. Even if the citation looks correct to you, it may have inaccuracies, either in the bibliographic information itself or in how the citation was formatted. Here is an example, including notes about missing or incorrect elements of the citation:
