This research guide is designed to help Vassar community members learn more about generative AI, with a focus on students who might use it in their classes. We created this because it can be hard to sort through the hype and sensationalization of the state of AI right now and find actual, reliable information. Vassar librarians have been working on learning what we can about how AI works, how it can be used, what some of the concerns with using it are, and what this all means for us here at Vassar. Though this is an evolving topic and we are still finding out more, we wanted to share what we've learned so far.

So what even is artificial intelligence? According to Encyclopedia Britannica, artificial intelligence is "the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings. The term is frequently applied to the project of developing systems endowed with the intellectual processes characteristic of humans, such as the ability to reason, discover meaning, generalize, or learn from past experience." There is a lot that falls under this category! Self-driving vehicles, Amazon's Alexa, Chat GPT, and the TikTok algorithm are all examples of how AI can be used.

People often use the term "AI" as a shorthand for generative AI tools like Chat GPT, Gemini (Bard), Perplexity, and others. These programs use what are called large language models to create outputs that are based in statistics about how words are most often paired together. This means that they are trained on huge datasets of written language and programmed to use that information to select the most likely combinations of words in response to the prompts that users give them. Sometimes people describe this as a more complex version of the predictive text that shows up on your phone or email.

We recommend the following resources to get a good overview of how generative AI works:

• Clarke, S., Milmo, D., & and Blight, G. (2023, November 1.) How AI chatbots like ChatGPT or Bard work – visual explainer. The Guardian.
• Nield, D. (2023, April 30.) How ChatGPT and Other LLMs Work—and Where They Could Go Next. WIRED..

There are a lot of potential uses of generative AI for coursework. Some examples of ways it could be used might include: creating an outline for a paper, generating keywords for searching, or editing writing to have a different style. As you can read about in the "AI at Vassar College" section of this guide, Vassar librarians are not the absolute authority on AI usage, but we are here to help you navigate these decisions as you make your way through your classes.

We do want to note a few important things that you should know about the different ways you might use AI:

We do not recommend asking generative AI to find books, articles, and other sources for your academic work. Generative AI tools like ChatGPT use probabilities to predict the most likely combinations of words in response to the prompts they're given, but are often not fact-checked. This means that they will often generate citations that sound real but do not actually exist. In addition, ChatGPT and many similar programs actually cannot tell you where they are getting their information from, because they are not set up to do so.

Some AI tools, such as Perplexity AI, do work more like a search engine and provide links to the actual pages that their answers come from. However, this does not guarantee that those sources will be reliable or relevant. If you do end up using AI tools to search for sources, it's important to verify that they are legitimate and that they are right for your research. See our guide on Evaluating Sources for tips.

Thank you to Naima Saini VC'23, Library Administrative Fellow in 2023 - 2024, for creating this guide.