The Lawyer-AI Sandwich: Defining the Human-in-the-Loop Architecture

The Lawyer-AI Sandwich: Defining the Human-in-the-Loop Architecture

How do you translate years of legal intuition into a digital product? In this episode of AI Practice Notes, host Michael Fahner talks with Marla Miller, a tax attorney and the Founder/CEO of 9to5 Legal Docs, about the transition from practitioner to builder. Marla shares her insights on why generic AI models often miss the mark in high-stakes legal work. They dive deep into the "Lawyer-AI Sandwich"—an architectural framework where the lawyer provides the initial context and the final strategic review, leaving the "messy middle" of drafting and data synthesis to the AI. Whether you're a skeptic or an early adopter, this conversation is all about how to stay "in the loop" while still capturing the massive efficiency gains of modern technology. Episode Highlights & Timestamps * [00:00] – Introduction: Meet Marla Miller, tax attorney, LLM, and CEO of 9to5 Legal Docs. * [02:15] – The Logic of Tax: Why a background in tax law is the perfect primer for building logical AI systems. * [06:45] – Beyond Generic: Why generic probability isn't enough for legal work and how attorney-vetted expertise changes the output. * [11:30] – The Lawyer-AI Sandwich: Defining the Human-in-the-Loop architecture—human context on the front end, AI in the middle, and human judgment on the back end. * [16:00] – The "Review" Myth: Reframing the pushback that "reviewing AI takes too long" and identifying where the true efficiency gains live. * [23:45] – The Future of Legalese: Will AI lead to more concise contracts or a flood of AI-generated text? * [29:30] – Mentorship in the AI Era: How junior associates can use AI to accelerate pattern recognition rather than just automating grunt work. * [35:30] – The Business of Law: A candid look at the rise of "AI-Native" venture-backed law firms vs. practitioner-led innovation. Key Quotes "AI will do what you want it to do... but it doesn't understand why that's actually bad for you in six months or a year down the road." — Marla Miller "The context window for a human lawyer is just bigger right now. We need to be the curators of the content." — Michael Fahner About Our Guest Marla Miller is a tax attorney and the Founder/CEO of 9to5 Legal Docs, a platform designed to provide attorney-vetted AI tools for startups and small firms. * Website: 9to5docs.com [https://9to5docs.com] * LinkedIn: Marla Calvert Miller [https://www.linkedin.com/in/marla--miller/] * Socials: @TheLawyerMarla (TikTok/Instagram) Connect with Socius Legal * Host: Michael Fahner * Website: Socius Legal [https://sociuslegal.com/] * LinkedIn: Michael Fahner [https://www.linkedin.com/in/michael-fahner-44374134/] Follow for more on the future of legal work.

Introducing AI Practice Notes by Socius Legal

What is the real impact of AI on the legal profession? Welcome to AI Practice Notes, a new podcast by Socius Legal. Hosted by corporate lawyer and builder Michael Fahner, this show is designed for the lawyers who are ready to move beyond the buzzwords and into the practical reality of Legal AI. In this series, we explore the intersection of law and technology, providing a grounded perspective on how Generative AI is reshaping daily practice—from AI contract drafting and due diligence to client communication and emerging ethical standards. Whether you are a solo practitioner, part of a boutique firm, or in-house counsel at a high-growth startup, this podcast is your field guide to the shift. The era of legal AI is just beginning. Let’s set the precedent together. Subscribe now so you never miss an episode.  📧 Contact: Michael@sociuslegal.com (Subject: AI Practice Notes) 🌐 Learn more: https://sociuslegal.com/ Connect with Socius Legal * Host: Michael Fahner * Website: Socius Legal [https://sociuslegal.com/] * LinkedIn: Michael Fahner [https://www.linkedin.com/in/michael-fahner-44374134/] Follow for more on the future of legal work.

AI in Legal Practice: Navigating Hallucination, Bias, and Ethical Duties with Anthony Hayes

In this insightful episode of AI Practice Notes, Michael Fahner, founder of Socius Legal, sits down with Anthony Hayes, a corporate lawyer and computer scientist turned lawyer. Anthony Hayes, the Associate Dean for Strategic Initiatives and Information Systems at Albany Law School and leader of the cybersecurity and data privacy practice at Dorf Nelson LLP, discusses the profound impact of AI on the legal profession. Building on Socius Legal's recent white paper, the conversation dissects the critical, non-abstract problems of AI in legal practice: hallucinations, bias, and omissions. Learn how these issues are shaping AI behavior, lawyer usage, and regulatory responses. Hayes provides a crucial perspective on the long-term, radical transformation of law by smart software, contrasting it with the short-term reality of AI's practical limitations and ethical risks for practicing lawyers. Key Takeaways:  ●       AI's Fundamental Flaw: AI hallucinations are not just a "quirky bug" but a "systemic risk," with one study showing Large Language Models (LLMs) hallucinating between 58% and 82% of the time on legal queries. ●       The "Quiet Danger": Beyond obvious hallucinations (like the Aviana Airlines case), the subtle risk is bias. AI inherits biases—including those related to socioeconomic status or race—from human case law and historical data, leading to the "automation of inequality". ●       Automation Bias: The human tendency to trust an answer given with confidence, even when wrong, leads to lawyers outsourcing their judgment to a machine that lacks independent judgment. ●       Ethical Duties Triggered: Using AI activates several Rules of Professional Conduct, including competence (Rule 1.1), communication (Rule 1.4), confidentiality (Rule 1.6), and candor to the tribunal (Rule 3.3). ●       The Future of Hiring: While AI won't replace senior or mid-level lawyers right now, partners will prioritize hiring associates who can effectively use AI tools over those who cannot. ●       The Confidentiality Risk: Never put client details into the general-use AI tools, as the data could be "memorized" by the AI, inadvertently revealing client secrets.  Guest Information:  ●       Anthony Hayes: Associate Dean for Strategic Initiatives and Information Systems at ALB Law School and leads the cyber security and data privacy practice at Dorf Nelson LLP.  Did this episode change how you think about AI in your practice? Share this episode with a colleague and leave us a review on your favorite podcast platform!   Subscribe to AI Practice Notes for more essential conversations shaping the future of legal work. Connect with Socius Legal * Host: Michael Fahner * Website: Socius Legal [https://sociuslegal.com/] * LinkedIn: Michael Fahner [https://www.linkedin.com/in/michael-fahner-44374134/] Follow for more on the future of legal work.