Choosing AI Tools

What is generative AI?

Generative AI, according to Adam Ziwe at MIT News, is "a machine-learning model that is trained to create new data, rather than making a prediction about a specific dataset. A generative AI system is one that learns to generate more objects that look like the data it was trained on." 

What are Large Language Models (LLMs)?

Large language models (LLMs for short) are, according to James Dunham at Georgetown University, "systems that learn patterns in language from being exposed to a lot of it." They create text responses that look like answers to questions you ask.  (Are the answers correct?  Sometimes...)

What are Large Multimodal Models (LMMs)?

Large multimodal models (LMMs for short) are models that can incorporate multiple types of input in their training set, some of which may include text, video, and images.  LMMs may also produce multiple types of output as responses to user queries.

There are many AI tools out there you can use.  All of them have lacks and there are things you need to think about before you use them.  For example:

Ethics of AI tools

• Ethical and responsible use of AI for students (California State University)
• Ethical AI for teaching and learning (Center for Teaching Innovation, Cornell University)

AI tools and misinformation

• AI chatbots will never stop hallucinating (Scientific American)
• ChatGPT and Fake Citations (Duke University Libraries)

Cost of AI

• The Billion-Dollar Price Tag of Building AI (Time Magazine)
• The Hidden Cost of AI (University of Wisconsin-Madison)

Environmental Price

• AI's climate impact goes beyond its emissions (Scientific American)
• California wildfires raise alarm on water-guzzling AI like ChatGPT (Fortune)

Generative AI can be useful for:

• brainstorming potential topics for a paper
• coming up with keywords related to a topic

Be careful of using generative AI to:

• recommend articles for your research.  The articles it gives you may not exist.  (Generative AI can create something that looks like an article citation but doesn't check to make sure its newly created citation is a real article.) 
• summarize articles or readings.  These summaries tend to be incomplete or lacking detail.
• write material for you.  Generative AI tools don't know what good writing looks like; they have many examples of good, bad, and mediocre writing in their databanks.

If you're using generative AI, always double-check the information it gives you against trusted information sources (books or articles on your subject, for instance).  If you're not sure how to find reliable information sources, ask your librarian!

There are many AI tools to choose from, not every tool will work for your task. How do you choose which one is the right one for your task?

• Is this tool a good match for my task?
  • Can it do what i need it to do?
• Is this tool easy to use?
  • Can I learn to use this tool without too much help?
• What is the quality of this AI's output?
  • Are the results accurate, relevant, complete, and useful?
• Where did the data come from?
  • What information was used to train the AI, and how might that affect how well it works or if it might have any biases?
• What are the ethical considerations of using this tool?
  • Are there moral or societal issues to think about when using this tool?
• What assistance/documentation is available?
  • Are there documentation, training resources, support, or guides?

The links below can help you evaluate AI tools.