Artificial Intelligence for Law and Finance

An open-source textbook bridging cutting-edge AI technology with practical applications in legal and financial domains. Designed for practitioners, regulators, academics, and technology builders working at the intersection of artificial intelligence, law, and finance.

Project Overview

Artificial Intelligence for Law and Finance: A Modern Textbook at the Intersection of AI, Law, and Finance addresses the growing need for comprehensive educational resources as AI transforms professional practice. The book provides both theoretical foundations and practical guidance for deploying AI systems responsibly in high-stakes domains.

Target Audience

The textbook serves four primary reader groups:

• Legal Professionals: Lawyers, compliance officers, and legal operations specialists
• Financial Practitioners: Analysts, risk managers, and financial advisors
• Academic Researchers: Graduate students and faculty in law, finance, and computer science
• Technology Builders: Entrepreneurs and engineers building AI solutions for regulated industries

Table of Contents

Part I: Foundations — LLMs and Prompting

1. LLM Primer and Mechanics — Understanding transformer architecture and model behavior
2. Conversations and Reasoning — Multi-turn interactions and chain-of-thought techniques
3. Structured Outputs and Tool Use — Function calling, JSON schemas, and external integrations
4. Multimodal Fundamentals — Vision, audio, and document understanding
5. Prompt Design, Evaluation, and Optimization — Systematic approaches to prompt engineering

Part II: Agents and Agentic Systems

6. What Is an Agent? — Conceptual primer and history of agents and agentic AI
7. How to Design an AI Agent — Architectures, protocols, and technical evaluation
8. How to Govern an Agent — Oversight frameworks, escalation protocols, and compliance

Part III: Knowledge Graphs & Semantic Web

9. Foundations for Law and Finance — Ontologies, taxonomies, and semantic modeling
10. Operations with LLMs — Graph-augmented retrieval and knowledge integration

Published Chapters

Individual chapters are released as working papers as they reach completion:

Chapter 6: What is an Agent? A Conceptual Primer and History of Agents and Agentic AI

• Authors: Michael J. Bommarito II, Jillian Bommarito, Daniel Martin Katz
• Published: November 2025
• Pages: 60
• Available on: SSRN

This chapter synthesizes nearly a century of scholarship across eight disciplines—philosophy, law, economics, cognitive science, and computer science—to propose a three-level hierarchy of agency:

• Level 1 (Agent): Goal, Perception, and Action (GPA)
• Level 2 (Agentic System): Adds Iteration, Adaptation, and Termination (IAT)
• Level 3 (Agentic AI): Fulfills all six properties using AI/LLMs for planning and orchestration

The chapter provides a practical evaluation rubric to help practitioners distinguish genuine agentic systems from sophisticated tools or single-shot chatbots, and explores critical dimensions such as the autonomy spectrum and entity frames.

Chapter 7: How to Design an AI Agent: Architectures, Protocols, and Technical Evaluation

• Authors: Michael J. Bommarito II, Daniel Martin Katz, Jillian Bommarito
• Published: December 2025
• Pages: 95
• Available on: SSRN

This chapter focuses on the architectural principles required to allow agents to function as cognitive work systems analogous to, and in concert with, professional teams. The analysis is organized around ten fundamental questions that shape an agent’s operational reality:

• Input Mechanisms: Triggers, intent, perception, and memory
• Execution Strategies: Planning, delegation, and action tools
• Safety Layers: Termination conditions, human escalation protocols, and governance

The chapter argues that robust design requires architectural literacy—a necessary bridge between technical implementation and professional obligation. Behind each design question lies a decision with real tradeoffs that determine what a system can do, how reliably it performs, and how it fails.

Chapter 8: Governing AI Agents: Risk, Compliance, and Accountability in Law and Finance

• Authors: Jillian Bommarito, Daniel Martin Katz, Michael J. Bommarito II
• Published: December 2025
• Pages: 69
• Available on: SSRN

This chapter proposes a governance framework for agentic AI systems in legal and financial services. Unlike passive AI tools, agentic systems can take consequential actions without human approval, creating risks that traditional compliance models often fail to address.

Key contributions include:

• Risk-Based Framework: Scales oversight requirements based on autonomy, operation duration, stakeholder interests, and objective setting
• Five-Layer Regulatory Stack: Encompasses foundational law, professional ethics, sector-specific regulation, AI-specific rules (including EU AI Act), and voluntary assurance standards
• Organizational Models: Evaluates centralized, federated, and embedded approaches with RACI matrices for responsibility assignment
• Professional Safeguards: Emphasizes human-in-the-loop and human-in-command architectures as critical for satisfying fiduciary and regulatory obligations

The chapter demonstrates how legal and financial institutions can adopt agentic AI while managing liability exposure and reputational risk through a maturity-based adoption path

Key Themes

Professional Safeguards

The book emphasizes responsible deployment in high-stakes domains:

• Attribution: Clear provenance for AI-generated content
• Escalation Protocols: When and how to involve human experts
• Confidentiality Controls: Protecting privileged information
• Audit Trails: Maintaining accountability in automated workflows

Autonomy Spectrum

Analysis of the continuum between:

• Delegated Proxies: AI systems acting under explicit human direction
• Self-Directed Entities: Systems with greater operational independence

Practical Evaluation

Rubrics to help practitioners distinguish:

• Genuine agentic systems from sophisticated tools
• Single-shot chatbots from iterative agents
• Marketing claims from technical capabilities

Open Source Commitment

The textbook is released under Creative Commons Attribution 4.0 International (CC-BY-4.0), enabling:

• Free access for students and practitioners worldwide
• Adaptation for classroom and training use
• Translation and localization efforts
• Community contributions and improvements

Resources

• Website: ai4lf.com (launching soon)
• GitHub Repository: ai-law-finance-book
• Full Draft PDF: Available in the repository
• Individual Chapter PDFs: Released as working papers

Impact

This project addresses critical gaps in professional education:

1. Definitional Clarity: Cutting through hype to establish rigorous frameworks for understanding AI capabilities
2. Cross-Disciplinary Synthesis: Bridging technical AI research with legal and financial practice
3. Practical Guidance: Moving beyond theory to actionable deployment strategies
4. Professional Standards: Establishing norms for responsible AI use in regulated industries

The book demonstrates that effective AI education for professionals requires both technical depth and domain expertise, setting a new standard for interdisciplinary AI resources.