Week 8: Section Activity

Activity 1: Each student will read aloud the passage they selected from Frankenstein and explain how they thought it was relevant to AI. After each student reads, the TF will ask if anyone else picked that same passage and launch a dialogue if several students chose the same passage. The TF will record on the board or screen the key ideas that everyone proposed connecting the novel with AI. After everyone has shared their pre-prepared ideas, you'll engage in a brainstorming discussion to suggest and explore connections beyond those already proposed. The final goal is a list of themes/ideas/connections, created by the individual and collective efforts of the section, recorded by the TF. In each case the connection should refer to a specific passage or quote from the novel.

Activity 2: GPT Challenge. Each student will try to use ChatGPT to see if it can come up with ideas that are new/creative beyond what the section proposed. Try at least 2 of the following modes of interaction with ChatGPT:

• Direct interaction with plain GPT-4, relying on its "memory" of the Frankenstein text
• Interacting with the custom GPT, 1188 Frankenstein Analysis Tool, which has the full text in its knowledge base (meaning it can search the text using contextual search based on vector embeddings of the novel)
• Uploading the text of the novel Download text of the novel as a file upload in plain GPT-4, which will allow it to search the text using Python tools but won't have the contextual vector embeddings for the novel.
• Copying and pasting the entire text of the novel Download text of the novel into the GPT-4 chat box, which means the entire text is contained in the "context window" for ChatGPT. You should probably preface the pasted text with some instructions for what you want GPT to do with it.

In each case, try to ask for ideas that connect Frankenstein with debates about AI with support from specific quotes or passages from the novel. Discuss your findings with your neighbors, and your TF will ask you to report out what you've uncovered. Do you notice differences in ChatGPT's interactions between these various modes?