We cannot go through a single day, it seems, without considering the burgeoning use of generative artificial intelligence (GenAI) and its impact on the practice of law. With TECHSHOW 2025 around the corner, where this will be a ubiquitous topic, it seemed to be a good time to hear from the “boots on the ground” about how practicing lawyers are using, or considering the use, of GenAI.
Please meet our virtual panel contributors:
Jeffrey O'Brien: Yes, I use GenAI daily, averaging 15 to 20 prompts or queries per day. My primary use case is organizing oral case notes. GenAI transforms my dictations about cases, strategies and document details into clear, legible content accessible to everyone on my team. I also frequently rely on tools like Perplexity for legal research. Perplexity leverages the RAG (Retrieval-Augmented Generation) model, which excels at generating accurate and well-cited responses, and I verify these citations for accuracy, significantly enhancing the efficiency of legal research. Additionally, to manage large volumes of documents, I use an OpenAI application programming interface (API) to generate one-sentence summaries of PDFs in bulk, saving significant time compared to manual uploads and processing.
One key challenge is understanding how different large language models (LLMs) interpret prompts, as each has unique processing quirks. Additionally, frequent model updates often alter response behaviors, requiring consistent monitoring and adaptation.
Rosanna Berardi: We've integrated AI across multiple aspects of our practice, from legal analysis to client communications. Beyond time savings, we've found AI particularly valuable for brainstorming complex issues and enhancing our written work. The technology has surprisingly helped identify analytical blind spots we might have missed. However, we maintain a strict verification protocol––all AI output undergoes thorough human review to ensure accuracy and reliability.
O'Brien: When it comes to confidentiality, there are two key paths to consider. The first path is for those using AI tools or any third-party software where data is stored externally—whether it’s an AI platform, Google Docs, Microsoft 365, etc. For these tools, it’s critical to treat them all with the same caution: understand their security measures, assess their access permissions and recognize their limitations. This ensures that client data remains as secure as possible, no matter the tool being used.
For those who want to take things to the next level or have the need to, the second path involves integrating AI directly into your own systems. In my office, we’ve built our platform to integrate AI tools directly within our case management system. This approach keeps sensitive information in one secure, unified location and eliminates the need to transfer data across multiple platforms.
Berardi: We take a two-pronged approach: using only enterprise-grade AI solutions with robust security features and implementing comprehensive redaction protocols before any client information enters an AI system. While awaiting specific AI guidance from our jurisdiction, we're applying existing confidentiality standards rigorously, treating AI platforms with the same caution as any third-party service provider.
Safiyya Vankalwala: The Federal Court of Canada has detailed when lawyers must disclose GenAI use in court filings. While Ontario courts have yet to adopt similar guidelines, other provincial courts have introduced AI disclosure requirements, and the legal community remains cautious after notable AI misuse cases.
The Federal Court of Canada released a Notice on the Use of Artificial Intelligence in Court Proceedings on May 7, 2024, detailing the requirements for the use of GenAI in court-submitted materials. The notice provides guidance on when you must declare you have used AI, specifically distinguishing between when content used in court submissions was provided by GenAI versus when GenAI “merely suggest changes, provide recommendations, or critique content already created by a human who could then consider and manually implement the changes.” Justice Alan Diner, in the video Compliance with the Notice of on the Use of Artificial Intelligence, states that the Court recognizes the benefits of GenAI but emphasizes that transparency enables the Court and other parties to mitigate its potential risks. Additionally, the Court’s Interim Principles and Guidelines on the Court’s Use of Artificial Intelligence states that GenAI will not be used by the Court for judicial decision making without consultation with the public and key stakeholders. Additionally, the Canadian Judicial Council has released Guidelines for the Use of Artificial Intelligence in Canadian Courts, providing Canadian judges with guidance on AI’s potential role in the legal system, as well as its benefits and risks.
Laura Keeler: In my jurisdiction, the Massachusetts Superior Court sanctioned a lawyer who submitted three separate legal memoranda that had four hallucinated case citations. The lawyer was fined $2,000. Smith v. Farwell; 12-007-24. This court opinion was picked up by several news outlets, including LawSitesBlog, in its aptly named “Not Again! Two More Cases, Just this Week, of Hallucinated Citations in Court Filings Leading to Sanctions.” The court sanction seems to have made some lawyers more wary, though situations like these are easily preventable; verifying cited cases for accuracy and relevance, as should always be done before submitting a court filing, will ferret out any hallucinations.
Berardi: After evaluating multiple platforms, we've settled on Claude AI for sophisticated legal writing and analysis, and Perplexity AI for research support. Each tool was selected for its specific strengths––Claude's nuanced understanding of legal concepts and Perplexity's ability to provide current, cited information.
O'Brien: For day-to-day operations, I rely on an AI system built into my Salesforce case management platform. It helps with summarizing documents, creating timelines and interacting with client data. For drafting and refining documents, I use ChatGPT’s Canvas feature. This tool allows for an iterative approach where I can make edits and receive real-time feedback to improve my work. For basic questions and guidance, I turn to ChatGPT and Perplexity to get quick answers and direction. The citations provided by these tools can be very valuable and are a much better resource than using Google for research.
Recently, I’ve started appreciating Claude by Anthropic for its more relatable and human-like interaction style, which feels more intuitive compared to other AI tools. Finally, I use the OpenAI API for large-volume tasks, such as summarizing PDF documents and interacting with my database. Unlike the other tools, it requires a more built-out approach rather than being plug-and-play, but if you’re willing to take this extra step, the number of things you can achieve with GenAI is endless.
Margeaux Green: Locally, the Washington State Bar Association’s Legal Technology Task Force is actively exploring the integration of AI’s role in law practice, ethical compliance and emerging risks. The task force’s recommendations are due in August 2025. Nationally, bar associations and task forces are issuing guidance to help lawyers navigate AI responsibly. The ABA and state bar associations have issued ethics opinions and guidance to help lawyers navigate AI use, including ABA Formal Opinion 512. Florida, New York, Michigan, California, Pennsylvania and New Jersey have also issued guidance on AI and ethical obligations.
O'Brien: The full impact of AI on the legal profession remains uncertain. However, the industry is undergoing a Darwinian shift where tech adaptation will be key to survival. Lawyers who fail to embrace AI risk falling behind, unable to match the speed, precision and cost-effectiveness of AI-integrated practices. These lawyers will likely become extinct. In contrast, I believe a new breed of lawyer is emerging—those who can harness AI as a collaborator. These forward thinkers will not only thrive and drive innovation, but I believe some will even expand the reach of legal services.
Brent Hoeft: Of course, we will be replaced! So, let's be nice to our future AI overlords! All jokes aside, I do not believe that GenAI will replace lawyers entirely, but I do think that it will replace many of the tasks that we currently look at as essential to “lawyering.” I believe GenAI will fundamentally change what it looks like to be a lawyer.
While GenAI excels at tasks such as drafting initial drafts of documents, conducting legal research and summarizing case law, it cannot replicate the nuanced judgment, ethical reasoning and strategic thinking that lawyers employ to represent their clients. Lawyers serve as advisors, advocates, negotiators and problem-solvers, roles that depend heavily on emotional intelligence, human relationships and the ability to navigate complex, often ambiguous real-world scenarios. GenAI, despite its power, lacks the understanding of context, moral considerations and interpersonal skills necessary to effectively represent a client’s interests or appear in court.
Instead of replacing lawyers, GenAI is more likely to become an integral tool that enhances their capabilities. By offloading routine and time-consuming tasks, GenAI allows lawyers to focus on higher-value work that requires lawyer expertise, compassion and creativity. This shift could lead to more efficient legal services, lower costs for clients, and greater access to justice for underserved populations.
The time saved by implementing GenAI for backend tasks could result in improved client service and satisfaction, as lawyers will have more availability to provide high-level legal services. There is no doubt that the lawyer’s role will evolve with the introduction of GenAI; those professionals who resist this change may find themselves at a disadvantage compared to those who adapt to and leverage GenAI in their practice.
As the saying goes, making predictions is difficult—especially about the future. If current trends continue, I believe we will see the augmentation and transformation of lawyering rather than the replacement or extinction of lawyers.
Berardi: AI will transform legal practice but won't replace lawyers. The profession fundamentally requires human judgment, emotional intelligence and ethical reasoning that AI cannot replicate. Instead, AI serves as an 'intellectual amplifier,' enhancing our ability to serve clients more effectively while maintaining the crucial human elements of legal practice.
Keeler: I have heard many iterations on the theme of “law practices that use AI will eclipse law practices that do not.” While I don’t foresee that GenAI will replace lawyers full-stop, I do think it’s a turning point for how professional services (legal and otherwise) are offered and priced, with alternative fee arrangements to increase and many first levels of tasks being augmented by AI. Just as prior tech progressions like document automation brought down the time of how long it takes to perform legal work, GenAI is far quicker than humans on first drafts, making likely to replace the activities historically done by law students and associates especially. (Activities such as drafting, reviewing, researching, transacting, recommending courses of actions.) In response, legal education and training of newer lawyers needs to evolve, to train more in critical thinking, advocacy, client relations, evolving technical competence and being more solution oriented.
If you want to learn more, Jordan Furlong has terrific ideas about reinventing the legal profession, building the lawyer of the future and developing forward-looking competencies.
Berardi: AI has proven invaluable for content creation and thought leadership, helping us translate complex legal concepts into accessible content for our blog and social media platforms. We use it to analyze trends, identify emerging issues and generate preliminary drafts––always applying our legal expertise to verify and customize the final output.
Green: One of my favorite uses for GenAI is its ability to help lawyers identify business management challenges and streamline workflows—skills that often feel out of reach, especially for solo and small firm practitioners balancing client work with running a business. GenAI acts like a business management assistant and helps lawyers clarify challenges, brainstorm solutions and automate tasks so they can focus on high-value, client-centered work.
AI can break down problems step-by-step, helping lawyers find practical solutions and improve efficiency.
A few examples of how lawyers can use GenAI to address business management challenges:
AI can assist in developing a structured timeline that reduces overwhelm and keeps deliverables on track.
“Write a friendly but professional email to remind a client about an overdue invoice.”
“Now draft a firm follow-up email for a client who hasn’t responded to the first notice.”
“Draft a step-by-step client onboarding checklist for my personal injury practice, including gathering medical records, signing the retainer and setting case expectations.”
“What are three ways I can speed up my client intake process while ensuring accuracy and compliance with RPCs?”
AI can generate actionable ideas, such as automating parts of the process or creating templates for faster follow-up.
I also think AI is an excellent technology training tool.
“Explain how to create and apply heading styles in Microsoft Word to automate section numbering and keep formatting consistent throughout a document.”
“Show me how to create reusable blocks of text for standard client update emails in Microsoft Outlook.”
“Explain how to set up a template in Word for fee agreements that automatically pulls in client details like names and dates.”
Jared Jaskot: AI has helped me improve my email beyond my wildest expectations. AI lets me strategically craft messaging. Rather than struggling to write an email, I tell the AI what my goals are and what the situation is and what tone I want. I also include prior correspondence. This exercise also helps me clarify what I want from the email in a way that is helpful. I then edit the email to my liking and find that the product is clearer and more thoughtful while avoiding needless provocation.
Most of my client conversations are in Spanish, I record, transcribe and translate the conversations. I then feed the transcript into AI with an instruction that transforms it into a case note for my practice management system. I have always struggled with taking good notes and now my notes are excellent thanks to transcription, translation and summarization. This saves time and adds a higher level of depth and detail than I ever believed I could achieve.
Keeler: Overall, I find AI to have fantastic potential! It is especially helpful at addressing the blank page issue. However, there are significant challenges, namely: 1. the black box problem: even developers aren’t quite sure how the algorithms function, and this must be solved before narrow AI progresses to general AI; 2. companies have changed the names, settings and terms of service without warning. Lawyers don’t want to repeatedly check a company isn’t training on their data, for example; 3. it’s challenging when people rely on AI as a catch-all crutch rather than a tool that still needs human reasoning and careful input. Iterations are important, for the first prompt and first output are rarely going to be sufficient; and 4. Since AI can produce convincing-sounding wording and can manipulate audio and video, I fear how this will distort people’s trust in the truth as persuasive disinformation can proliferate.
As tech progresses and it becomes harder to tell when text/audio/visuals are manipulated, I foresee that as a coming challenge for trusting the authenticity of evidence. I recommend following people like Judge Scott Schlegel, who are on the frontlines of thinking about how to modernize courts and consider AI in the justice system.
Green: One of the most challenging aspects of getting started with GenAI is prompt writing. This is a skill that’s essential for getting useful results but feels very new. Writing a good prompt isn’t like asking Google a quick question. With AI, the more specific, detailed and intentional you are, the better the response.
Tips for drafting an effective prompt:
1. Define the format. If you need a list, an outline, multiple options or specific examples, say so. You might try, “provide three versions of a client email: one formal, one friendly and one concise,” or “Summarize this information into a bullet-point list with action items.”
2. Refine through iteration. If the first result isn’t perfect, refine the prompt with clearer instructions. Your process might look something like this:
o First prompt: “Create an email to remind clients to pay outstanding invoices.”
o Follow-up: “Make the tone friendlier and add a closing line offering to answer any questions about the invoice.”
3. Test and experiment. The more you practice, the better your prompts will get. Start small and low stakes. For example, use AI to brainstorm meal plans, create packing lists for a road trip or organize a weekly schedule. Low-pressure tasks like these give you a feel for how AI works and what kinds of refinements improve the result.
Prompt writing is a skill that takes practice, but it’s worth the effort. A well-crafted prompt turns AI into an effective assistant—helping you brainstorm, organize tasks and reduce time spent on repetitive work. By learning how to give clear, structured instructions, you’ll get better results and find more value in using AI day to day.
Green: GenAI is changing how legal services are delivered, and this will inevitably affect what clients consider “reasonable” fees. By streamlining repetitive tasks—like formatting documents, organizing emails or drafting starting points—AI allows lawyers to work more efficiently, focusing their time on high-value work like strategy, client relationships and problem-solving.
These efficiency gains will likely push firms to reconsider traditional hourly billing, as clients increasingly expect cost savings for tasks that AI can help complete quickly. Flat fees or task-based billing models may become more common, offering transparency while aligning with the benefits of AI-driven workflows.
At the same time, initiatives like the Washington State Bar Association’s Pilot Test of Entity Regulation reflect a broader trend toward using technology to improve access to justice. By lowering the cost of legal services, AI has the potential to make legal help more affordable and accessible—particularly for underserved communities.
For lawyers, the challenge is balancing these efficiencies with ethical obligations: ensuring they leverage AI responsibly, maintain competence, and deliver real value to clients. Thoughtful use of AI can help lawyers stay competitive, meet client expectations and better serve those who need their expertise.
Hoeft: “Death of the billable hour!”—a prophecy that numerous times has been foretold. The billable hour is deeply ingrained in the practice of law. Do I believe GenAI will be the end of the billable hour? No, I think it will persist in specific practice areas. However, a significant shift in how lawyers bill is very likely. The key question is what constitutes a reasonable fee in connection with GenAI and the time-saving efficiency gains it offers in delivering legal services.
Those who know me might be surprised that I am not fully convinced GenAI will ultimately lead to the demise of the billable hour. Spend enough time with me and you are likely to hear my rants about how the billable hour can be a disservice to clients, promoting inefficiency and waste on the part of attorneys and firms, while also creating a chilling effect on open communication with clients, who might hesitate to share additional questions or information if they worry about the cost of each one-tenth of an hour billed. If you aim to create happy clients and a positive perception of your law firm and the legal profession the billable hour is not going to get you there. (I will step down from my soapbox now.)
Nina Carney’s article “Formal Opinion 512 and the Reasonableness of Fees When Using AI” in the December issue of the ABA’s Law Practice Today, used an example I often use when speaking to lawyers about GenAI, the evolution of legal research from books, to databases on computers, to online legal research platforms. The jump in efficiency from book-based, library legal research to online tools was massive. Yet, the billable hour did not disappear then, despite predictions, and law firm revenues did not plummet, even with these tremendous efficiency gains.
That said, it is essential to consider what factors should be considered when determining reasonable fees. Given that GenAI’s efficiency gains can be applied across many tasks in legal practice, it's logical that the time spent should not be the primary factor in assessing the reasonableness of a fee. Rather the ultimate value delivered to the client should be the primary focus of whether a fee is reasonable.
As a result, alternative billing models such as value-based billing, phased billing approaches, fixed-fee billing, outcomes-based billing or a hybrid billing model are likely to become more common as law firms and clients navigate the impacts of GenAI on the legal profession.
Green: Lawyers and law firms must approach Gen AI with clear systems and policies to ensure compliance with the Rules of Professional Conduct. Trust alone isn’t enough. Safeguards must be built into every stage of AI use to protect client confidentiality, maintain competence and uphold ethical obligations.
Steps firms should take include:
Vankalwala: Caution and compliance. With the various directives and widespread coverage of hallucination cases, lawyers have become increasingly vigilant in their use of GenAI. A notable instance in British Columbia, where AI generated fictitious cases that a lawyer used for court submissions, is often part of formal and informal conversations amongst the legal community.
It is expected that lawyers will continue exploring the use of GenAI. However, the widespread news of the challenges it presents is likely to encourage lawyers to develop their own personal and firm-wide guidelines for usage, including adherence to mandatory requirements and directives.
