AI & Society
[Unedited image created by AI. Errors maintained for demonstration purposes.]
Overview
This section covers the societal implications of AI, emphasizing the importance of ethics and the current uses of ChatGPT and GenAI. It covers notable headlines, such as the controversy with Sports Illustrated’s use of AI-generated articles and fake profiles, and explores AI’s impact on productivity, job displacement, and creative processes. Additionally, it introduces the concept of prompt engineering and its significance in optimizing AI outputs, and provides further reading suggestions for those interested in deepening their understanding of AI’s effects on various aspects of society. Finally, it demonstrates how to use text-based generation tools such as ChatGPT and features examples of using the voice interface.
Videos
AI & Society
Generative Text Part 1
Socrates Bot
Generative Text Part 2
Mock Job Interview
Suggested Readings:
Suggested Assignment:
Assignment Description: Exploring Prompt Engineering with LLMs
Overview:
Objectives:
- Gain practical experience in designing prompts for LLMs.
- Understand the impact of prompt design on the responses generated by AI.
- Compare and contrast the effectiveness of different LLMs in understanding and responding to prompts.
Instructions:
- Select Two LLMs: Choose two LLMs from the following list: OpenAI’s ChatGPT, Anthropic’s Claude or any others from this list.
- Develop Prompts: Create at least three unique prompts that you can try on each LLM. Refer to the ChatGPT cheat sheet and the Anthropic Prompt library in the readings for help in crafting prompts. These prompts should be designed to test the model’s ability to understand and generate relevant and coherent responses. Consider varying the complexity and specificity of your prompts. Try iterating your prompts to get better reponses as you go. Consider prompts that simulate real-world scenarios where AI might be used in communication and collaboration. This can involve customer service interactions, team meetings, or negotiations where AI tools provide support or automation.
- Document Responses: Record the responses from each model to your prompts. For ChatGPT you can share a link to the chat. For others you may need to take screenshots or copy and paste the text. Note any significant differences in how the models handle the same prompt.
- Analysis: Write a 500-word analysis comparing the performance of the two LLMs. Discuss which model performed better and hypothesize why certain prompts worked well or poorly with each model.
Submission Requirements:
- A document containing your prompts, the responses from the LLMs, and your analysis.
- Format your submission as a PDF.
- Include screenshots or direct text outputs from your interactions with the LLMs.
Rubric for Prompt Engineering Assignment
| Criteria | Excellent (90-100%) | Good (80-89%) | Satisfactory (70-79%) | Needs Improvement (<70%) |
|---|---|---|---|---|
| Prompt Creativity | Prompts are highly creative and effectively test different capabilities of the LLMs. | Prompts are creative and have a clear purpose. | Prompts are somewhat repetitive and lack clear objectives. | Prompts are not effective or are too simplistic. |
| Quality of Analysis | Provides a deep, insightful comparison of the LLMs with detailed explanations supported by specific examples from the responses. | Analysis is well-reasoned with some specific examples. | Analysis covers basic observations without much detail. | Lacks depth or critical analysis of the LLM responses. |
| Clarity and Organization | Submission is exceptionally well-organized, with clear documentation of prompts and responses; analysis is coherent and logically structured. | Well-organized submission and analysis with minor clarity issues. | Organization is adequate, but some parts may be confusing or poorly structured. | Poor organization and lack of clear structure in documentation and analysis. |
| Adherence to Submission Guidelines | Fully adheres to all submission requirements and guidelines. | Mostly adheres with minor deviations from guidelines. | Meets the basic requirements but misses some elements. | Fails to meet multiple submission guidelines. |
Header Image by J.J. Sylvia IV using MidJourney is licensed under a Creative Commons Attribution Non-Commercial Share Alike (CC BY-NC-SA) 4.0 International License
