Skip to Main Content

TH 600 Research and Writing for Theology: Artifical Intelligence

Artificial Intelligence

Large language models of Artificial Intelligence are trained on massive amounts of text data from books, articles, and websites. They produce human-like answers to questions and prompts. Claud.ai and ChatGPT.com are two examples.

Google's NotebookLM is a mini-large-language-model. It may be used for notetaking, organizing, and summarizing.

Undermind.ai is a search engine for complex research questions. It often requires the user to refine the prompts multiple times. It uses AI-powered search algorithms to search hundreds of papers, analyze the results, and adjust its search strategy based on concepts not simply keywords. It provides both narrative and analytical results.

Some databases such as ProQuest and JSTOR provide AI Research Assistants within the result screen to help you quickly grasp the focus of the article, chapter, or book, and assist in determining the value of the resource in your research.

Using AI Assistance

  • You may request a list of potential keywords to use in searching academic databases.
  • You may request help in brainstorming ideas of research topics.
  • You may request a list of scholarly databases related to your topic.

Cautions in Using AI Assistance

  • Privacy - Do not provide information to an open Large Language Model that you would not publish on the internet. There are significant questions about how the data is used, stored, and protected.
  • Accuracy - Sometimes programs provide biased results or make up information and citations. Verify the results using trusted academic sources.
  • Copyright and Licensing - Providing an article for the program to summarize may violate copyright or licensing agreements with the database providers. 
  • Academic Integrity - Write your own papers.
  • Environmental Impact - Large Language Models use significant amounts of water and energy and produce significant amounts of wastewater and emissions (https://www.cetjournal.it/cet/23/107/018.pdf). A study released in July of 2024 found the emissions produced to train ChatGPT3 equivalent to 600 flights between London and New York (https://doi.org/10.48550/arXiv.2408.01453).