# Parsing a reading in the sciences


prompt used: "parsing a scientific journal article"

## description
Students can use a generative AI tool to produce quick summaries of a reading, as a form of reference material. Students can submit an article and ask for a summary of the piece, specifying pertinent themes, concepts, and other content the student deems necessary or helpful. Using this summary, the student could prepare for class discussion by formulating questions on their own or even using the AI tool to help develop questions.

Depending on the course goals, this type of engagement with AI might give students too much content, or it might mitigate against their deep engagement with course readings. You might want to set guidelines for appropriate or inappropriate use.

## activity
1. Log into chatGPT using the Harvard AI Sandbox
2. From search bar at the bottom of the page, specify that you'd like to use GPT 4.
* We recommend selecting GPT-4 32K for the best results, especially with long prompts
4. Hit Shift + Enter to type questions below the quoted text.
* If you only hit Enter or Return, this will automatically submit your prompt.
* You can also add three ticks (''') to indicate a paragraph break if you prefer.
5. Ask chatGPT to provide you with a summary that includes:
* a paraphrase of the article's methods
6. Wait for the interface to give you a response.
* This may take a moment.
* You may then copy and paste the response for your own use or reference.
7. Once you have a response from chatGPT, you can ask it to help you come up with a follow-up experiment you could run in response to this paper's findings.

---
## Transcript of activity example:
* ```USER: I would like your help understanding the methods these authors used in their study. Can you help me understand the methods of this study?```

* link to the article we used: [A Programmable Dual-RNA–Guided DNA Endonuclease in Adaptive Bacterial Immunity](https://www.science.org/doi/10.1126/science.1225829), MARTIN JINEK, KRZYSZTOF CHYLINSKI, INES FONFARA, MICHAEL HAUER, JENNIFER A. DOUDNA , AND EMMANUELLE CHARPENTIER. We downloaded the methods section and input this into GPT 4 32K.
* AI:


![alt text](https://files.slack.com/files-pri/T0HTW3H0V-F0636F3AC2X/screen_shot_2023-10-27_at_3.14.45_pm.png?pub_secret=9f514dc30e)