By OLGA V. MACK
July 31, 2023
Anybody can interact with a generative AI tool. Getting the results you want, however, is an art unto itself.
In a recent conversation among a group of in-house lawyers, one GC shared how she’d used ChatGPT to write a data processing agreement. “Just 25 prompts later, I had an optimized DPA. I’d saved so much time and effort,” she said.
A skeptic countered, “But ChatGPT doesn’t know anything beyond 2021, and so much has changed with privacy laws since then.”
The GC explained, “That’s part of what the 25 prompts were about. I started with a good DPA template. Then, during a dialogue with ChatGPT, I tweaked the agreement to include company specifics and recent privacy law changes from the past two years. It was well worth it.”
Anybody can interact with a generative AI tool. Getting the results you want, however, is an art unto itself — one we all need to master. Here’s how you, the experienced and trained legal operator, can use the chain-of-thought process to guide conversations with AI language models in ways that help you explore and connect ideas for more productive outcomes.
AI Language Models Lack Critical Thinking Skills
Though AI language models generate text that is often indistinguishable from human writing, they can’t interpret context reliably or understand language nuances. AI-generated text may sound convincing and well-reasoned, but AI models don’t fully grasp their underlying meaning or implications. They need us to fill in the gaps.
AI language models are trained to pinpoint patterns and probabilities in large datasets not to make value judgments or to take a stance on a particular issue. Because of their training, they reproduce existing patterns and ideas rather than genuinely novel or innovative thoughts.
By relying solely on AI-generated text, a lawyer might easily overlook important details or make incorrect assumptions. That’s because AI tools lack our human critical thinking skills.
Analytical Thinking Is A Crucial Component Of The Legal Skillset
In the chain-of-thought process, you inject critical thinking skills into an AI tool’s output, and AI brings creativity to yours. Critical thinking helps you identify underlying assumptions, dissect premises and conclusions, and evaluate the strength and consistency of your and others’ reasoning.
Human lawyers break down complex situations into smaller parts and systematically examine each aspect to reach rational conclusions. Assessing different possibilities, identifying contradictions, and establishing connections are all essential skills for effective legal practice.
Through the chain-of-thought process, you systematically introduce critical thinking and legal reasoning into your interactions with generative AI tools. As a result, the tools can tailor more practical and relevant responses.
Here are four tips for engaging in more productive conversations with AI models using the chain-of-thought process:
Define Clear Goals And Provide Context
You must give the AI model a well-defined context and clear goals to get responses tailored to your specific legal needs. What do you want to achieve? What information does the tool need to achieve this?
Identify key concepts or arguments ahead of time and prepare to examine each individually. This helps you keep the conversation on track.
Break down complex ideas into smaller components for discussion.
Begin the conversation with a clear and focused question or prompt.
Include examples and illustrations of the topic at hand.
Offer clarifications, explanations, and additional details as needed to help the AI model understand the situation at hand and your goals.
Make Connections Among Various Ideas AI Presents
AI language models identify patterns and connections in large text datasets. In turn, you must do the same when analyzing their output. Does it match up with what you know about the legal issue? Is it consistent with what you have seen elsewhere?
Look for common themes or ideas that emerge. What works together? What is the same or similar across different sources?
Stay open to new insights. Don’t hesitate to challenge your assumptions!
Examine the output from different angles and viewpoints, and consider alternative interpretations of the data. Ask AI to do the same.
Always use your legal expertise and judgment to determine whether output is accurate and up to date. Double-check sources and cross-reference with outside documents to verify the information.
Structure Follow-Up Questions And Prompts Logically
Remember, you’re in an ongoing dialogue with AI in which you must bring logic and clarity.
Ask targeted follow-up questions.
Request real-world examples.
Seek clarifications, explanations, and additional details to ensure that you understand the output.
Be specific and frame prompts in a well-reasoned manner to get refined responses that lead to more productive discussions.
Continuously Refine Your Interactions With AI Models
Continue to refine your interaction techniques. AI language models constantly learn from new data inputs, fine-tuning their understanding and increasing proficiency in various areas. Subtle adjustments in your approach can yield better results over time.
Combining the output of AI language models with your critical thinking skills and legal expertise leads to high-quality legal work that takes all relevant factors into account. AI interaction strategies like the chain-of-thought process can help unlock new levels of productivity and responsiveness that propel your law career forward in exciting ways.