In coding and technology circles, a "playground" or "sandbox" refers to a secure, isolated environment for experimenting with new tools and techniques. The metaphors are apt: these are spaces for us to play around without worrying about getting perfect results. When you're first testing out the technology, think of your interactions as playgrounding. See what you get and don't worry if what you get isn't good right away. We can also think about generative AI as a playground - a space to explore innovative teaching methods without risk. Use this guide to not only help you explore the technology but also to explore ideas for the classroom. 

Key features of this AI sandbox include:

1. Risk-free environment: Experiment freely without impacting live systems.

2. Immediate feedback: Observe results instantly to refine your approach.

3. Minimal setup required: Most AI tools are readily accessible via web browsers.

4. Hands-on learning: Align with pedagogical theories emphasizing active engagement (Piaget, 1963).

___________________

To start exploring, try this out: 

Copy the following prompt into your preferred AI tool (e.g., ChatGPT):

"Act as a 1L student, draft a case brief for Palsgraf v. Long Island Railroad Co."

Analyze the output:

• How accurate is the content?

• Is the level of detail appropriate?

• What strengths or weaknesses do you notice?

______________________

By treating generative AI as a playground, we can systematically explore its capabilities and limitations, enhancing our instructional strategies in the process.

We encourage you to experiment regularly and share your findings with colleagues. Together, we can leverage this technology to advance legal writing instruction.

Prompting in the Context of AI

When working with Artificial Intelligence (AI), particularly language models and chatbots, "prompting" refers to the initial input or question posed to the machine to generate a specific output or answer. This input activates the pre-trained algorithms, which then search, process, and formulate a suitable response based on the data they have been trained on. Essentially, a prompt acts as an interface between human queries and machine-processed answers.

Language models like GPT (Generative Pre-trained Transformer) use a type of machine learning called Transformer architecture. These algorithms analyze the prompt based on their training data and generate a sequence of text in response. The model considers the syntactic and semantic context of the prompt to produce a coherent and contextually appropriate output. For more technical applications, specialized models trained in specific domains, such as law or medicine, provide more accurate and domain-specific responses. 

Prompting plays a crucial role in educational applications of AI. The effectiveness of machine-generated responses can significantly impact the learning experience. Crafting well-designed prompts is key to eliciting detailed, nuanced, or critical responses from the AI. These responses can foster constructive classroom discussions or aid in research and academic writing. Moreover, the sophistication of the prompt can be adjusted according to the students' or researchers' needs, making it a customizable learning tool.

Prompting may be an art, but it's not foreign to anyone trained in writing. It's as easy as accounting for the rhetorical situation. Just be clear about the author's voice, the intended audience, the genre, the subject, and the purpose of the writing you're eliciting. See this article for more. 

Iterative Prompting

Iterative prompting is an advanced method of interacting with AI systems, particularly language models. It involves a sequence of refined or adjusted queries to elicit more precise or contextually appropriate responses. Unlike one-off prompting, where a single query is posed, iterative prompting is a dynamic, ongoing process that adapts to the user's needs or the unfolding conversational context.

The core idea of iterative prompting is to use the output from an initial prompt as the basis for subsequent queries, fine-tuning the AI's responses over multiple iterations. This approach helps users zero in on the information or analysis they seek, improving the clarity and relevance of machine-generated content. Iterative prompting is particularly useful for complex or multi-faceted questions where a single prompt may not suffice to generate a thorough or nuanced answer. If the initial output is too broad or misses the point, a follow-up prompt can clarify or narrow the focus.

In educational and academic settings, such as law, iterative prompting can be an effective tool for research, hypothesis testing, and drafting complex arguments or documents. For example, a legal scholar might begin with a broad query about a legal principle, refine it based on the AI's response, and continue this process until a highly focused and relevant body of information or analysis is produced. This iterative approach complements traditional research methods and can be a powerful supplement to resources like legal databases or scholarly articles (See Susskind, R. "Tomorrow's Lawyers: An Introduction to Your Future" (2013) for a discussion on technology in legal research).

Iterative prompting involves a sequence of queries designed to refine and focus the output from an AI system, which can be particularly useful for complex tasks such as generating the discussion section of a legal memo. Below is an example that demonstrates how one might use iterative prompting to accomplish this task.

Iteration 1: Initial Query

Prompt: "Generate a discussion section for a legal memo on the implications of the 'fair use' doctrine in copyright law."

AI Response: "The 'fair use' doctrine in copyright law permits the unauthorized use of copyrighted material under specific conditions such as criticism, comment, news reporting, teaching, and research. However, the application of this doctrine varies case by case, making it a complex legal issue."

Iteration 2: Refining for Depth

Prompt: "Provide a more in-depth analysis of the factors considered by courts in applying the 'fair use' doctrine in copyright cases."

AI Response: "Courts typically evaluate four primary factors when applying the 'fair use' doctrine: the purpose and character of the use, the nature of the copyrighted work, the amount and substantiality of the portion taken, and the effect of the use upon the potential market for the copyrighted work. Each of these factors carries its own set of sub-considerations and contributes differently to the final judgment."

Iteration 3: Seeking Specific Examples

Prompt: "Include case law that serves as precedent for each of the four factors of 'fair use' in copyright law."

AI Response: "In the realm of 'fair use,' several landmark cases establish precedent for each of the four primary factors. For instance, 'Campbell v. Acuff-Rose Music, 510 U.S. 569 (1994)' set a precedent for the first factor by ruling that commercial nature alone does not negate fair use, emphasizing transformative use. The second factor was clarified in 'Harper & Row v. Nation Enterprises, 471 U.S. 539 (1985)', which held that unpublished works rarely fall under fair use. For the third factor, 'Kelly v. Arriba Soft Corporation, 336 F.3d 811 (9th Cir. 2003)' determined that using thumbnails in search engines could be considered a small and insubstantial portion. Finally, the fourth factor was exemplified in 'Sony Corp. of America v. Universal City Studios, 464 U.S. 417 (1984)', which stated that home videotaping did not have an adverse effect on the potential market for the copyrighted television shows."

Iteration 4: Summarizing for Memo Format

Prompt: "Summarize the in-depth analysis and case law into a concise discussion suitable for a legal memo."

AI Response: "The 'fair use' doctrine in copyright law remains a complex yet pivotal aspect, governed by four primary factors. Courts weigh the purpose and character of use, often leaning towards transformative uses as exemplified by 'Campbell v. Acuff-Rose Music.' The nature of the copyrighted work also plays a role, with unpublished works rarely falling under fair use, as clarified in 'Harper & Row v. Nation Enterprises.' Additionally, the amount and substantiality of the portion taken are considered, as demonstrated in 'Kelly v. Arriba Soft Corporation.' Finally, the impact of the use on the potential market for the original work is a critical factor, with cases like 'Sony Corp. of America v. Universal City Studios' providing guidance. Each factor is a nuanced criterion, shaped by precedent, and carries significant implications for any fair use analysis."

Through this iterative process, the AI's response has evolved from a broad overview to a concise, yet in-depth discussion section suitable for a legal memo, complete with case law for support. Of course, you don't want to trust that it's right about the case law or the analysis, but do note how it will continue to expand as you continue to follow-up. 

Below is an example of how one might proceed through multiple iterations to refine the structure and style to align with CREAC.

Iteration 5: Introducing CREAC Structure

Prompt: "Rework the earlier discussion summary on 'fair use' doctrine into a CREAC format, starting with the Conclusion section."

AI Response: "In summary, the application of the 'fair use' doctrine in copyright law is a multifaceted issue, shaped by a four-factor analysis as delineated by precedent."

Iteration 6: Building the Rule Section

Prompt: "Following the Conclusion, add the Rule section to explain the general legal principles that govern the 'fair use' doctrine."

AI Response: "The governing rule for 'fair use' stipulates four central factors: the purpose and character of the use, the nature of the copyrighted work, the amount and substantiality of the portion taken, and the effect of the use on the potential market for the copyrighted work."

Iteration 7: Adding Explanation and Application

Prompt: "Continue with the Explanation and Application sections, elaborating on how the rule applies to specific cases. Include the precedent case law discussed earlier."

AI Response: "In interpreting these rules, courts consider a range of precedents. For instance, 'Campbell v. Acuff-Rose Music' elaborates on the first factor, emphasizing transformative uses. 'Harper & Row v. Nation Enterprises' clarifies that unpublished works are generally less likely to fall under 'fair use,' speaking to the second factor. The third factor is elucidated in 'Kelly v. Arriba Soft Corporation,' which found the use of thumbnails permissible as they represent a small and insubstantial portion of the work. Finally, the impact on the market, the fourth factor, was explored in 'Sony Corp. of America v. Universal City Studios,' concluding that home videotaping did not adversely affect the market for the copyrighted shows."

Iteration 8: Closing with a Revised Conclusion

Prompt: "Conclude the CREAC-formatted discussion by summarizing how these rules and precedents collectively shape the application of the 'fair use' doctrine in copyright law."

AI Response: "In light of the four governing factors and the specific case law that provides interpretation, it's evident that the 'fair use' doctrine is a nuanced legal concept. Each case is subject to a unique, multi-pronged evaluation that considers not only the letter of the law but also the interpretive latitude granted by judicial precedent."

Through these iterative prompts, the AI's response has been refined to fit the CREAC paradigm.