How Lawyers Can Use Generative AI for Legal Design – The Basics

Lawyers can use generative AI for legal design to enhance efficiency and effectiveness. The technology offers enormous potential for simplifying legal language, contract analysis, predictive analysis, and streamlining drafting. Generative AI refers to artificial intelligence models that generate new content, such as text, images, or audio, based on the training data they have been exposed to.

Legal design centers on the user and is grounded in empathy to deliver solutions that effectively meet their needs. Generative AI enables lawyers to experiment, generate novel ideas, and explore faster ways to produce user-centric solutions.  

This advanced language-processing AI generates human-like responses by understanding context and language nuances, providing meaningful and relevant outputs based on its training on vast human-written text datasets.

Generative AI is a powerful tool for legal design, but it is unable to replace the nuanced understanding, empathy, and emotional intelligence that design-thinking lawyers bring. It’s a tool that helps lawyers simplify contracts, analyze, and ideate, allowing them to brainstorm and have fun exploring potential solutions.

Lawyers must use generative AI responsibly and be cognizant of ethical considerations, privacy, and copyright issues. Machine models augment but are unable to replace critical thinking, human judgment, or legal expertise. (Yet?)

Lawyers interested in the basics of generative AI and legal design are able to start by using the technology to simplify legal language and contracts, and the article explores how to simplify contracts as a use case. Prompts for generative AI agents are included to get you started.

What is Generative AI in Legal Design

Generative AI in Legal Design allows lawyers to use models and develop prompts to produce words and documents that are explicitly designed for the user. It is an inexpensive way for lawyers to ideate, redraft, explore language, produce clauses, contracts, and visuals, and deliver effective solutions that clients understand. Legal design thinking ensures generative AI is always used to produce the best outcome for the user.

Generative AI in legal design is not limited to contract design and simplification, but it is an excellent place to start.

Generative AI is a kind of artificial intelligence that produces content such as text, images, and audio by replying to a prompt. The generative AI models discussed in the article refer to tools such as ChatGPT by OpenAI, Claude by Anthropic, Gemini by Google, and Copilot by Microsoft.

Generative AI models understand semantics and how humans interpret and use language. The models understand that people don’t use the same words to mean the same thing, which offers vast opportunities in legal design.

Why can generative AI help with Contract Simplification?

Generative AI can help with contract simplification because the models are trained on enormous datasets, allowing them to predict language patterns, understand context and semantics, and generate outputs based on human language and words. Contracts contain words, and generative AI understands words and is an ideal tool in the legal design toolkit.  Note: The models are far from perfect and always require human oversight. 

The features listed below allow lawyers to use legal design thinking to simplify contracts.

• Natural Language Understanding: Generative AI understands human language. It comprehends the words, context, and nuances of what’s being said. Generative AI understands the complex words and contexts in legal documents. It allows lawyers to harness this understanding to simplify legal language.
• Language Generation: generative AI produces meaningful text. The models create human-readable content that conveys information clearly and straightforwardly. Generative AI simplifies dense legal language, replaces it with more straightforward legal language, and makes it more accessible to a wider audience.
• Contextual Responses: generative AI understands conversational context. The output it generates is coherent and relevant (most of the time). Maintaining context throughout a conversation adds authenticity to our interactions with generative AI.

Generative AI understands what lawyers are saying and responds contextually and sensibly. Interactions with the models are natural, meaningful, and ultimately more helpful. Generative AI has improved from robotic chatbot days, but we must remember we are interacting with machines, not humans.

Looking for an example of generative AI simplifying contracts and legal language? I asked Claude to give me an example of complex legal language rewritten in simple terms, and the output is below.

“Complex Legal Phrase 

“In the event of a material breach of this Agreement by either Party, the non-breaching Party shall have the right to terminate this Agreement upon giving thirty (30) days’ prior written notice to the breaching Party, during which period the breaching Party may cure such breach.”

Simplified by generative AI

“If one side fails to follow an important part of this contract, the other side can end the contract. However, they must give 30 days’ written notice first. During that 30-day period, the side that failed to follow the contract has a chance to fix their mistake.”

In this example, the generative AI has rephrased the legal jargon into more accessible language:

“Material breach” becomes “fails to follow an important part”

“Terminate this Agreement” is simplified to “end the contract”

“Non-breaching Party” and “breaching Party” are replaced with more straightforward terms like “one side” and “the other side”

“Prior written notice” is rephrased as “written notice first”

“Cure such breach” is explained as “fix their mistake”

An experienced lawyer familiar with legal design must review the output for accuracy and assess the output regarding the user. Still, it’s a fast and low-risk way to ideate and continue prompting the model to refine the words to best suit the user.

How Legal Design Can Help with Generative AI

Legal design can help with generative AI by ensuring that lawyers using the tools always focus on the users and put the user at the center of the problem-solving process. Legal design provides lawyers with a point of reference when interacting with the models because it is human-centric at its core. 

Simplifying contracts using Legal Design and generative AI means incorporating the user perspective into the prompts and every subsequent decision. 

For example, when simplifying a contract for freelance services, the freelancers who commit to the contracts may have a different understanding of legal concepts and implications than someone experienced in dealing with complex contracts. Legal design ensures generative AI knows the goal is to simplify the contract language so that a freelancer can easily understand it.

When can Lawyers use Generative AI with Contracts?

Lawyers can use generative AI with contracts to break through writers’ block when simplifying contract language, creating contract clauses, summarizing and explaining contracts, and getting suggestions for revisions. 

Generative AI can make contract drafting efficient and enjoyable. Writing contracts is daunting, and generative AI is not just a tool; it’s a creative companion. It’s not always right, but it can challenge your thinking and open new avenues for you. 

Lawyers can use generative AI with contracts in the ways listed below.

1. Breaking through writer’s block. We’ve all experienced staring at a blank page, struggling to find the right words. Generative AI serves as a muse for overcoming writer’s block. Generative AI is ideal for lawyers who are at a loss for words or seek simple ways to express something. It provides fresh ideas, alternative phrasings, and inspiration to get your creative juices flowing. Lawyers often become stuck in their language and need more ideas because we were taught to write contracts a particular way. Simplification feels like learning a new language, and this is where generative AI can offer assistance.

2. Generating contract text. Generating sections of contracts becomes easier and faster with the help of generative AI, eliminating the need to go through piles of precedents sitting in databases manually. This capability saves invaluable time. Generative AI assists in crafting clear and concise language, whether dealing with boilerplate clauses or specific terms. Always remember that AI-generated content is a starting point for review and refinement rather than being directly employed without oversight.

3. Easy summaries and explanations. Distilling a lengthy contract into a concise summary becomes easier with the help of generative AI. It can extract key points and explanations, ensuring the audience understands the essence without getting lost in the details. Summaries and explanations are particularly useful for highlighting the most critical aspects of complex agreements, making them more accessible and understandable.

4. Suggestions for revisions. Use generative AI to make revision suggestions if you’re looking for ways to improve your contract. It’s like having an AI editor that reviews your work and suggests enhancements. This assistance is particularly beneficial in identifying areas that may require further clarification or simplification, and it can also suggest alternative phrasing to make a contract more understandable and user-friendly.

What are the Risks and Concerns Related to Generative AI in Law?

The risks and concerns relating to generative AI in law include accuracy and reliability, overreliance on the technology and lack of human oversight, ethical and privacy considerations, and copyright and ownership. 

The risks and concerns related to generative AI in law are listed below.

• Accuracy and reliability: Generative AI outputs include hallucinations or mistakes, and around 20% of outputs are incorrect, depending on the model. Fact-check every output to ensure it complies with the law in your area and check for bias and fairness. The models are trained on data that contain societal biases and unfair assumptions. Legal design thinkers are fair and unbiased.  “Treat GenAI as a modern, very lazy first-year associate. You know you must look after the associate because they aren’t ready to draft contracts without guidance and supervision.” Aaron Crews in LegalWeek 2023.
• Ethical and privacy considerations: Using generative AI to produce legal content raises questions about transparency, accountability, fairness, and privacy. Ensuring that your generative AI usage aligns with ethical standards and safeguards sensitive data is crucial. Ask yourself: Where does the data you input into generative AI go? Who has access to it, and for what purposes? Understanding how your data is used is essential to maintaining control and ensuring compliance with data protection regulations.
• Copyright and ownership: The copyright and ownership of generated data are unclear. Relying strictly on generative AI outputs increases the likelihood of inaccuracy, which is why lawyers must apply human oversight. Stay abreast of legal developments on copyright and ownership, ensure compliance, and always customize output. 
• Overreliance on the technology: Blindly accepting outputs is dangerous because they’re often inaccurate. Legal design requires lawyers to ensure the user is at the center of the solution. Every user is different, and every output must be tailored to the user’s needs. Lawyers are smart cookies; don’t let AI replace your brain. 

Golden rule: be thoughtful and diligent

Always remember a fundamental principle—be thoughtful. Don’t rely solely on a single source of information, even if it’s as advanced as generative AI. Always review, validate, and cross-reference the results against established legal standards and expertise and double-check for unfair assumptions or bias.

How to Start Using Generative AI for Contract Simplification

To start using generative AI for contract simplification, follow the steps below.

1. Select Relevant Text and Prepare Inputs. Begin by selecting the portions of your contract text that you’d like to simplify or make more user-friendly. Remove client-specific data or confidential information before submitting the text to GenAI to ensure the privacy and security of sensitive information. Stealing prompts from generative AI models is occurring. Cornell University developed a “prompt stealing reconstructor” that was able to “reconstruct the original prompts based on the generated answers and the extracted features” in the paper Prompt Stealing Attacks Against Large Language Models, 2024. Ensure your prompts do not contain confidential information.
2. Define the User for the Output. Identify the end user of the simplified contract and what their struggle with the text is. Understanding your audience—whether clients, colleagues, or a broader public—helps tailor the language and complexity of generative AI’s output to their needs and comprehension levels.
3. Provide Clear Prompts and Context. Give clear prompts and context. Let the AI know what specific tasks you want it to perform—whether it’s simplifying complex language, rephrasing intricate clauses, or making legal terms more accessible. Clearly outlining your expectations helps the machine understand your goals and generate relevant suggestions.
4. Review and Iterate. Generative AI’s output is a starting point, and exercising your legal judgment is essential. Iterate through the suggestions, consider their impact on the legal accuracy and meaning, and refine them as needed. You control the final output, ensuring it aligns with your professional standards.

How to Write Prompts for Better Legal Generative AI Outputs

To write prompts for better generative AI outputs, follow the steps below.

1. Be specific and clear.  Prompts that communicate the task generative AI must do are essential. Avoid vague or ambiguous language that leads to incorrect or irrelevant responses. Be specific about what the prompt is focusing on, whether it’s a sentence, a paragraph, or a particular question related to the contract section you want to simplify.
2. Conversation Context. If your interaction involves a series of back-and-forths, include the conversation history to provide context. This helps generative AI understand the ongoing discussion.
3. Iterative Approach. Have multiple rounds of interaction to refine the output. Each time you provide feedback or ask for further simplifications, you build upon the previous responses.

How can Lawyers Communicate Effectively with Generative AI?

Lawyers can communicate effectively with generative AI by following the steps below.

1. Ask for Explanations. If the AI-generated output needs to be clarified or more accurate, ask GenAI to explain its reasoning. This can help you understand the model’s thought process and guide it more effectively.
2. Provide Feedback. If the output doesn’t meet your needs, provide constructive feedback. Specify what needs improvement and suggest alternative phrasing if possible.
3. Refine in Steps. Start with broader simplifications and then narrow down as needed. This gradual approach can help the AI model produce more accurate and contextually appropriate results.
4. Balance Simplicity and Legality. While you aim for simplified language, remember that legal accuracy and validity must be maintained. If an output sacrifices legal precision, it might not be suitable.

Examples of User-Centric Generative AI Prompts for Contract Simplification

Examples of user-centric generative AI prompts for contract simplification are listed below. 

1. Example Contract Simplification Prompts

Prompt 1: “Our company uses a non-compete restriction clause in the employment contracts of our key employees. We get a lot of questions from the employees about the clause and need to simplify the language. They worry about their career progression.  

Please simplify the following clause in the employment contract regarding non-compete restrictions.

Use plain language that is easily understandable for the employees worried about their future job opportunities.

Don’t change the meaning of the clause.”

Prompt 2: “I work in a big multinational company with very complex and long sourcing contracts. The challenge is that some of our suppliers are very small companies who don’t have their own counsel to read and explain the contracts. We especially have challenges with making the suppliers understand the requirements for sustainability, such as compliance with our codes of social responsibility. We have decided to simplify the content to ensure better compliance. 

Please reframe the compliance clause using clear, concise language that preserves its legal integrity. 

Emphasize simplicity to help small suppliers in understanding and adhering to our sustainability standards.” 

Prompt 3: ”I work in a law firm and many of our clients are in the publishing industry, where copyright and other intellectual property rights are at the core of the business. Such rights are defined and transferred by contracts. However, the writers and illustrators have a hard time understanding their rights and obligations when they sign a publishing contract. 

Please redraft the IP clause to demystify it for writers and illustrators, addressing their common questions. 

Include practical tips for presenting these concepts in a way that’s both enlightening and engaging.”

2. Example Contract Explanation Prompts

Prompt 1: “I am trying to explain the content of this contract, but I’m out of ideas. I need a fresh look at it. I’m looking for a casual tone, as the users of this contract are young influencers, who use a very different language and expressions from traditional legal language.

Please explain the content of this contract as if I was a cool 16-year-old teenager, adopting a casual tone that maintains the essence of the agreement while making it accessible and engaging.”

Prompt 2: “My client, a tech startup founder, finds the concept of share vesting in the shareholders agreement challenging, especially with concerns about VC control. They have no background in law. 

Please explain the vesting of shares using metaphors and simple language that resonate with the entrepreneurial spirit. 

Aim for clarity and relatability to ease the founder’s concerns.”

Prompt 3: ”I received a complex M&A contract from a large law firm representing the potential investor. The content is challenging for me to comprehend. As a lawyer from a civil law country, I’m not accustomed to the lengthy and intricate contracts typical of common law countries. 

Please break down the indemnification clause into simpler terms, focusing on its implications for my business. 

Provide a clear, easy-to-understand explanation that bridges the gap between civil and common law perspectives.”

3. Example Contract Revision Prompts

Prompt 1: “I am negotiating a vendor agreement and we go back and forth with it as it’s lacking proper language on privacy. The users of this vendor agreement are procurement managers, who handle a volume of these agreements, and would appreciate simplified contracts to expedite the review process. 

Please review this vendor agreement and suggest revisions to ensure that it complies with relevant data privacy laws and regulations under GDPR. 

Aim for brevity and clarity to support procurement managers in efficient reviews.”

Prompt 2: ”I advise a labor union in contentious discussions regarding the collective bargaining agreement. The union is very firm about the conditions for an annual leave, as is the employer’s association. Consequently, the negotiations are stuck. 

Please propose 10 alternative revisions to the annual leave clause that consider both parties’ stances while prioritizing the union’s key concerns. Strive for equitable solutions that facilitate breakthrough in negotiations.“

Prompt 3:”I am involved in finalizing a service contract with a technology provider. The main issue is the service level agreement (SLA) section, which the client sees as unbalanced and not providing sufficient coverage for them. The primary users of this contract are IT managers and procurement specialists, who deal with numerous contracts like this and prefer them to be straightforward to facilitate quicker decision-making.

Please analyze and suggest adjustments to the SLA section to ensure it’s balanced, clear, and actionable. Focus on defining obligations, performance metrics, and penalties in a manner that’s easily grasped by both parties.”

The more you interact with it, the more the models will learn your style and expectations, but the output never replaces your legal brain. Always review for accuracy.

How to Review Generative AI Output for Accuracy

To review generative AI output for accuracy, follow the steps below.

1. Fact-check and cross-reference. Carefully verify any factual claims, legal citations, or references from the generative AI against authoritative, up-to-date sources. Cross-check the information with reliable legal databases, case law, and statutory sources.

2. Review for consistency and coherence. Ensure that the generated output is internally consistent and logically coherent. Check for any contradictions, gaps in reasoning, or abrupt shifts in tone or context.

3. Assess for bias and fairness. Be mindful of potential biases in the training data or the model itself. Evaluate the output for discriminatory language, stereotyping, or unfair treatment of individuals or groups.

4. Scrutinize legal accuracy and validity. Carefully review the generated output for its legal accuracy, ensuring that it adheres to relevant laws, regulations, and legal precedents. Pay close attention to nuances in language and terminology that could have significant legal implications.

5. Involve subject matter experts. Collaborate with your fellow lawyers and subject matter experts to review and validate the generative AI output, especially in specialized areas of law or complex legal matters.

6. Conduct side-by-side comparisons. Compare the generative AI output with existing legal documents, precedents, or human-written examples to identify any significant deviations or areas of concern.

7. Iterative refinement. Provide feedback and refine the generative AI output iteratively until it meets the desired level of quality and accuracy.

8. Document and track changes. Maintain a clear audit trail of the changes made to the generative AI output, documenting the rationale behind each modification for future reference and accountability.

9. Establish quality control processes.  Implement quality control processes and checklists specific to the legal domain to ensure that the generative AI output meets the required accuracy, completeness, and legal validity standards.

10. Continuously monitor and update. Stay informed about the latest developments, limitations, and best practices related to generative AI in the legal field, and continuously update your evaluation processes accordingly.

Humans will always need humans. Generative AI is not a lawyer replacement in the context of legal design. Legal design is about people and making the law accessible to a wider audience. We can use emerging technologies to help us get curious and explore opportunities quickly, easily, and meaningfully. Ultimately, our capacity to empathize with and understand people allows us to design prompts that ensure generative AI delivers useful outputs.

It is not the other way around 🙂