Generative Artificial Intelligence (GAI) is a powerful, albeit decisive, tool. If you are aware of its advantages and its limitations and drawbacks, then you can use it to effectively streamline your research process without crossing ethical boundaries or committing plagiarism.
Before even starting to think about how to use AI in research: Talk to your professor! Every professor has a different stance on AI and it's use in the classroom; at the WCSU Libraries we encourage open communication throughout the scholarly process. It's important to be on the same page as your professor and give proper credit (especially in regards to creating citations) for the resources and tools you use in your research and creative process.
It might be even more complicated than you think, but until you have an effective understanding of how these generative tools "think" you'll never be able to use them effectively for research or personal use. Watch the short video below from the Massachusetts Library System (MLS) on how Generative AI tools work:
Before knowing what AI can do to help streamline your research process, it's vital to be aware of its limitations.
Creating False Information: Also known as "Hallucinations," generative AI tools may make make false or inaccurate statements when creating a response to one of your prompts. OpenAI even has a disclaimer in their FAQ about how truthful ChatGPT can be: Does ChatGPT tell the truth?
Harmful or Biased Responses: The sources of information used to train AI models come from all over the Internet, and containt the views and language reflected there. Using ChatGPT as an example once again, we can see that there's another disclaimer under Is ChatGPT Biased? In the disclaimer it states that it "skews towards Western views," that the their model "can reinforce a user's biases over the course of interaction," and that these "biases can harm students if not considered when using the model for student feedback."
For more information, watch MLS's video on the limitations of generative AI:
To get a full understanding of the ethical and legal concerns surrounding the creation, training, and use of AI it is highly recommended that you go straight to MLS's artificial intelligence guide. On their page Concerns About AI they list numerous articles, news sources, and research surrounding different areas of ethical and legal concerns. From general concerns, to copyright, to economic and labor impacts, this guide covers it. We also recommend you watch the video below (also made by MLS), which gives an excellent snapshot of these various concerns.
For students and faculty at WCSU one of the biggest concerns to have in mind when using AI is academic Honesty.
Academic Honesty: The improper use and citation of AI tools for your assignments may result in a violation of WCSU's Academic Honesty Policy. It is very important to be familiar with what the university constitutes as cheating or plagiarism, especially for the use of AI tools. Any time that you do use them, it's key, for your academic success, to be open with your professor of how and where you used GAI in your research process and how to properly cite that contribution in your work. MLA, APA, and Chicago Citation Manuals have all come out with updated guidelines for citing GAI in your papers.
Like the other research-oriented AI tools listed here, SciSpace is centered around using the power of AI to summarize and simplify academic research. While other AI tools, like Google's Gemini, have options to upload PDFs and "talk" with them, SciSpace was one of the first to implement this feature.
Using its "Chat with PDF" function, you can upload scholarly research and ask SciSpace questions about what's within. For example, if you upload a paper with a study, intervention, or controlled trial within it, you can ask SciSpace to summarize the methodology used by the researchers.
To see SciSpace in action check out the video below:
Perplexity is another AI tool that might help with your research process. One of the biggest "strengths" that Perplexity has to offer is its ability to cite its sources. When you ask it a question you'll see a number at the end of any statements that it's making, clicking on that number will take you to the sources that Perplexity is using to give you that information! You can even "filter" for where Perplexity will pull information from.
However, it is still important to take any information it presents you with a grain of salt, so to speak. Any answer you receive from an AI tool should be treated with a healthy amount of skepticism, even tools that cite their sources like Perplexity. At the end of the day, you're still relying on this tool to be able to reliably read, process, and present information to you; there is a chance it might present the information to you incorrectly, or the source that it is pulling from is incorrect or even worse, misinformation.
A good way to use Perplexity then, is as a backup for finding popular articles or sources of information. When beginning a project, it's always important to conduct some basic research, both in popular media (magazines, newspapers, etc.) and in academic media (scholarly journals, articles, encyclopedias, etc.), if you're having trouble producing results on Google, you can try asking Perplexity to find popular sources of information for you. As just mentioned above: ALWAYS take time to evaluate the sources it presents you for legitimacy before incorporating them into your research projects.
What sets ResearchRabbit apart from other AI-powered research tools is its ability to act as a "literature map." Using a starting, or a "seed," article, it then searches for relevant content and articles to link, displaying that content as a bubble map of connected subjects. While an account is required to access this tool, it's core functionality is free to use.
To see ResearchRabbit's literature map in action, watch the video below:
Like Perplexity, Consensus is another tool that will cite the information that it gathers and presents to you. What sets Consensus apart is that it searches through over 200 million research papers across scientific domains. On paper, it is a tool that allows you to ask questions and get a "consensus" from the scientific community. In practice, your mileage may vary based on what version/subscription tier you're using.
For example, using Consensus without an account will still give you cited answers, but most of the information contained within those sources will be locked behind a wall. Using a free account will get you access, but you're limited in the number of analysis tools that you can use.
While it is one of the best "research assistant" AI tools, for the purposes of your tenure at WCSU, it might not be worth the cost when you can make use of the same research assistance at the WCSU Libraries' Reference Desk for free.
Like Consensus, this tool is "free to start," wherein creating an account is required to use the tool for a limited time until a paid subscription is required to continue.
Elicit sources it's papers from Semantic Scholar on a weekly basis, their current library exceeding 126 million papers. More information can be found here: About Elicit
While it is a great research assistant tool, for the purposes of your tenure at WCSU, it might not be worth the cost when you can make use of the same research assistance at the WCSU Libraries' Reference Desk for free.
Below is a great video from Brandon University, in Canada, on how Elicit can be used to write a literature review. This video is great at highlighting both the strengths and weaknesses of this tool, as well as what you need to watch out for when writing literature reviews using AI.
ChatGPT is an evolving conversation toolset, meaning that it's great at maintaining consistency across long conversations. Keeping in mind it's struggles with bias and creating inaccurate information, it can still be an effective tool early in your research.
Keeping in mind that it's strength lies in conversation, try bouncing your early project ideas off of it. If you're not sure about your thesis topic, try using ChatGPT to brainstorm project ideas. If you already have your topic in mind, try using ChatGPT to synchesize your ideas into a thesis question, like the prompt below:
Example: Can you help me develop a research question about heart disease prevalence in adults over forty in the US geographic south?
Just be aware, sometimes the answers it gives below might be too general for your research projects. Try narrowing down the scope yourself before feeding it more details about what you want to research.
Another good way to use it is for developing your search strategy. A search strategy is the method, or steps, that you'll take to locate scholarly information for your research projects. This includes the places that you'll search for information (databases, library catalogs, etc.) and how you'll search them, including the keywords and Boolean logic that you'll need in order to do so.
Example: Can you help me come up with a search strategy for scholarly articles on heart disease prevalence in adults over forty in the US?
In the examples above we're using ChatGPT to refine and narrow down our ideas, rather than come up with them for us. Using GAI effectively and ethically in research means using it as a tool to enhance our research skills, not replace them!
The videos below give good examples of how to effectively format prompts for ChatGPT.
