Background to this exercise:
Search engines are just search engines, increasingly trained to provide you with the information they have been trained to show you based on your previous search history. LLMs (Large Language models) powered by Artificial Intelligence operate differently – they will provide you with answers that are based on probabilistic answers – they are not scraping the internet, they are putting together answers that it is learning to understand (like a 3 year old with a huge processing power).
Between these two computational modes of digital inquiry there is a lot of room to get wrong answers. Some of them are intentionally wrong – like fake news websites, others are wrong because the humans who are training the AI or establishing the guardrails of its training are embedding value judgements (see the recent Gemini debacleLinks to an external site.), and because of the way AI works, it will make things up.
But AI can be great – it can help you organize your sources chronologically, and even make sure you are citing them correctly. It can help you write formulas for excel and code in any language.
https://www.npr.org/2024/03/18/1239107313/google-races-to-find-a-solution-after-ai-generator-gemini-misses-the-mark
PROMPT: this exercise is to produce a piece of research and demonstrate how you know what you know about it. You will start with AI, then move to primary and secondary sources and keep track of this process and share it in your storymap.
For this assignment, we want you to use the LLM model of your choice to ask as much as possible about an event in history that you are interested in. Go ahead – give it a go. Think of any historical question.
Whatever question you ask, keep track of the LLM’s answers [take screenshots]. Then confirm the answers using steps you have learnt in class.
You have been using sources assigned in class. You also want to use Mike Caulfield’s Web Literacy for Student Fact-CheckersLinks to an external site. in module 1. The summary of his advice/technique is: 1) check for previous work, 2) go upstream 3) read laterally 4) circle back. Each of these steps is detailed in the e-book, you don’t have to go through each substep – but you do want to take these 4 steps seriously because they will protect you from making bad calls.
suggestions:
-possible LLMs: Google’s Gemini, Open AI’s ChatGPT, Anthropic’s Claude – it does not matter to me which one you use
-take screenshots of these steps so you have visuals to include in your storymap
-repeat the exercise: you may need to rephrase the question or ask follow up questions. In order to complete this exercise, you will have to ask multiple questions (remember that each question leads to more questions). If your first question leaves you with no more questions – ask a new question. You should have at least 3 related questions.
-confirm the answers. Use Mike Caulfield’s Web Literacy for Student Fact-CheckersLinks to an external site. to practice safe web-searching and avoid common pitfalls
-use all this information to present your research in a storymap, and use all the skills you have learned to make this the best possible expression of your research savvy. Your questions will have led you to a thesis statement – make sure it is explicit. Your presentation is a formal piece of research that is illustrated with images and other media, so make sure you are writing in full sentences and paragraphs and include each step of the research in the storymap.
