Skip to Main Content

AI & Research

Use this guide to learn how to use AI ethically and effectively to enhance your learning.

Each course has unique goals and outcomes, including how you use AI. Check with your professor and syllabus for specific information on how to appropriately use AI in your work.

Fact Check AI

It is your responsibility to check the information that AI gives you. AI will generate the most likely response, which can include common myths or factually incorrect information.

For example, when I asked Claude to give me a list of books by Sharon Ketcham, it generated one correct title, a book written by another author, and a plausible-sounding title that doesn't exist.. 

Prompt: Give me a list of books written by Sharon Ketcham.

Claude.ai: Sharon Ketcham has written several books in the field of practical theology and youth ministry. Here are her published works:

  1. "Reciprocal Church: Becoming a Community Where Faith Flourishes Beyond High School" (2018)

  2. "Faith Formation in a Secular Age" (2020)

  3. "Becoming Whole: Adolescent Discipleship in a Fragmented Culture" (2014)

Note that while I've tried to be accurate, I recommend double-checking this information as there may be more recent publications or works I'm not aware of since my knowledge cutoff is October 2024.

While some Generative AI tools are linked to the internet, they will still sometimes create sources that don't exist. In this example, the first peer-reviewed source is not only not peer-reviewed, but it doesn't exist! 

Prompt: Can you give me 10 scholarly sources on oceanic noise pollution?

Gemini: "Anthropogenic noise in the oceans: a threat to marine mammals" by C. Erbe et al. (2016) - This comprehensive review article in Marine Pollution Bulletin provides an overview of the sources of anthropogenic noise in the ocean and its impacts on marine mammals, including behavioral changes, physiological stress, and hearing damage."

AI & Bias

AI can imitate other people's bias. For example, when asked to generate images of students studying outdoors and in a library, ChatGPT generated images of students that are largely white, which is not an accurate racial makeup of US college students.

Prompt: Can you generate an image of college students studying outdoors?

ChatGPT:

A ChatGPT generated image of college students studying outdoors, which are predominately white.

Prompt: Can you generate another image of college students, this time studying in an old library?

ChatGPT:

A Chat-GPT generated image of largely-white students studying in a library.

Limitations of AI

Some have said that generative AI is the smartest, dumbest friend you can have. Here are five key limitations to be aware of while you work with your new smart dumb friend:

  • AI can produce and defend false information
  • AI is imitating patterns, not creating truly new content or analysis.
  • AI can imitate other peoples' bias. For example, if you ask an image generator to create a picture of a college student, they will generate white students.
  • Some AI large language models cannot access new information past a specific date (ChatGPT, Claude).
  • When you ask for sources, AI will often generate non-existent sources or citations.

Additional Resources on AI Limitations