Modern Artificial Intelligence for Legal Professionals
A working mental model of artificial intelligence for legal work: what machine learning, generative AI, language models, retrieval, agents and document AI actually do, taught through everyday law-office tasks, so you can tell what a tool is doing before you trust it with a matter.
-
What Artificial Intelligence Is (and Is Not) in Legal Work
40 min
-
Inside a Large Language Model: Tokens, Training and Prediction
45 min
-
Retrieval, Agents and Tools: The Systems Behind Legal AI Products
45 min
-
A Field Guide to Legal AI Tool Categories
40 min
What AI Can and Cannot Reliably Do
Why a fluent AI answer can still be wrong, how fabricated citations and stale law arise, why the same question can get two answers, and how to build the habit of doubt that keeps AI-assisted legal work verifiable. Taught with measured error rates, real sanctions decisions and annotated fictional examples.
-
Why an AI Answer Can Sound Like a Lawyer and Still Be Wrong
40 min
-
Hallucinations, Fabricated Citations and Stale Law
45 min
-
Context Limits, Ambiguity, Prompt Sensitivity and Model Differences
45 min
-
Automation Bias and the Discipline of Doubt
45 min
Confidentiality, Privacy and Attorney-Client Privilege
Learn what the duty of confidentiality actually covers, where an AI tool puts the text you give it, how consumer and enterprise accounts differ on training and retention as published on a stated date, how privilege and work-product protection can be put at risk when a third-party model is in the room, and a recorded decision method for asking "can I put this into this system?" before you type.
-
What Counts as Confidential, and Where AI Tools Put It
40 min
-
Consumer Versus Enterprise AI Accounts, and What Vendors Do With Your Data
45 min
-
Privilege and Work Product When a Third-Party Model Is in the Room
45 min
-
"Can I Put This Into This AI System?" A Decision Method
50 min
Professional Responsibility and AI
Apply the duties of competence, communication, fees, candor and supervision to generative AI in daily practice, then build and keep a record of what your own licensing jurisdiction and your forum actually require, because the answers differ from state to state and from judge to judge.
-
Competence and the Duty to Understand the Tool
40 min
-
Communication, Consent, Fees and Candor
45 min
-
Supervision of Lawyers, Staff and Software
40 min
-
The Jurisdictional Map: Checking Your Own Rules
50 min
AI-Assisted Legal Research
Run legal research with AI the way a careful lawyer runs it: a fixed eight-stage workflow from question to work product, a clear line between an assistant and an authority, a verification and citator step for every citation, and a provenance log the file can defend.
-
The Safe Research Workflow From Question to Work Product
40 min
-
Assistant, Not Authority: What AI Research Tools Actually Do
40 min
-
Citation Verification and the Citator Step
45 min
-
Documenting Research Provenance for the File
35 min
Prompt Engineering for Legal Professionals
A nine-part architecture for legal prompts, a recorded method for testing and improving them, and ten working prompt patterns for summaries, chronologies, issue spotting, deposition preparation, contracts, discovery, research, memos, policy analysis and client communication, each built so that a licensed attorney can verify the output quickly.
-
The Nine-Part Legal Prompt Architecture
40 min
-
Iterative Prompting: Refine, Test and Compare
40 min
-
Prompt Patterns for Summaries, Chronologies and Issue Spotting
45 min
-
Prompt Patterns for Contracts, Discovery, Memos and Client Communication
45 min
Drafting and Reviewing Legal Documents with AI
Hands-on drafting and review with AI across memoranda, correspondence, internal policies, contracts, demand letters, discovery papers and deposition outlines, applying one chain every time: AI draft, lawyer review, authority verification, factual verification, revision, approval, with the record that proves it was followed.
-
First Drafts Without First Mistakes
40 min
-
Redlining, Clause Analysis and Contract Comparison
45 min
-
Demand Letters, Discovery Documents, Deposition Outlines and the Approval Record
45 min
Evidence, Discovery and Large Document Sets
Hands-on practice running classification, extraction, chronology, summarisation, contradiction-finding and deposition preparation across a large fictional document set, with the privilege, metadata, custody and validation controls that keep an AI-assisted review defensible.
-
Classification, Extraction and Chronology at Scale
45 min
-
Summaries, Contradictions and Deposition Preparation
45 min
-
Privilege, Metadata, Custody and the Limits of Automated Review
50 min
AI for Law-Firm Operations
Apply AI to the operational side of a practice, from intake and scheduling through knowledge management, matter summaries, billing narratives, marketing and FAQs, with the intake, supervision, advertising and fee rules that govern each use and a ranked risk table that separates the routine automations from the ones that need a lawyer to sign off.
-
Intake, Scheduling and Client-Facing Automation
45 min
-
Knowledge Management, Internal Search and Matter Summaries
45 min
-
Billing Narratives, Marketing, FAQs and Ranking Use Cases by Risk
50 min
Building Repeatable Legal AI Workflows
Turn one-off AI chats into controlled, repeatable processes. Learn the seven-stage workflow (input, classification, AI task, verification, human approval, output, audit record), draw six common legal tasks through it with risk levels and control points, and design approval gates and audit records that people will actually use.
-
From Chatbot to Controlled Process: The Seven-Stage Workflow
40 min
-
Six Worked Workflows From Intake to Knowledge Search
50 min
-
The Workflow Canvas, Approvals and Audit Records
45 min
AI Agents and Autonomous Systems in Law
How AI agents differ from chatbots once they can call tools, browse, read files, send email and execute actions; a risk ladder from read-only to external action; the human-in-the-loop control points, least-privilege identities, logs and kill switches a firm needs; and how prompt injection turns an agent against its owner.
-
What an Agent Is: Tool Calling, Browsers, Files and Email
40 min
-
Risk Levels for Agents, From Read-Only to External Action
45 min
-
Human-in-the-Loop Control Points and Kill Switches
45 min
-
Prompt Injection and Other Ways Agents Are Turned Against You
45 min
Cybersecurity, Data Governance and Vendor Evaluation
Learn to read an AI vendor's security and privacy claims the way a reviewing attorney reads a contract: what SOC 2, ISO/IEC 27001 and 42001 and the NIST frameworks actually prove, how data storage, residency, training, retention and deletion really work, what identity, logging and API controls to demand, and how subprocessor, breach-notification and exit terms protect client information.
-
How to Read a Vendor's Security and Privacy Claims
45 min
-
Data Storage, Residency, Training, Retention and Deletion
50 min
-
Identity, Access, Logging and API Security
45 min
-
Contracts, Subprocessors, Breach Notification and Exit
55 min
Law-Firm AI Governance
Why a law office needs a written AI policy, the eleven sections it must contain, and how to draft clauses on approved tools, prohibited data, acceptable use, verification, disclosure, oversight, supervision, records, incidents, vendors and training that people can actually follow. The learner assembles a basic policy for a fictional firm.
-
Why a Firm Needs an AI Policy and What It Must Cover
40 min
-
Approved Tools, Prohibited Data and Acceptable Use
45 min
-
Verification, Disclosure, Supervision, Records, Incidents and Review
50 min
Implementing AI in a Legal Organization
Treat AI adoption as a business process rather than a purchase: start from a measured problem, map the workflow, classify the risk, run a chartered pilot with paired speed and quality metrics and human controls, then read the results honestly and decide to expand, modify or stop.
-
Problem First: Mapping the Current Workflow and Classifying Risk
40 min
-
Pilots, Metrics and Human Controls
45 min
-
Measuring Results and Deciding to Expand or Stop
45 min
Practical Legal AI Lab
Eight hands-on exercises on fictional material: audit an AI intake, mark up a flawed research memo, expose fabricated citations, rebuild a weak prompt, check an AI contract analysis, design a secure workflow, grade a vendor questionnaire and draft a firm AI-use policy.
-
Lab: Intake Analysis and the Flawed Research Memo
55 min
-
Lab: Find the Fabricated Citation and Fix the Weak Prompt
55 min
-
Lab: Review an AI Contract Analysis and Design a Secure Workflow
60 min
-
Lab: Evaluate a Vendor and Build an AI-Use Policy
60 min
Capstone: The Legal AI Implementation Plan
Build the sixteen-section Legal AI Implementation Plan a managing partner or general counsel could adopt: one organisation, one workflow, a measured baseline, a labelled risk and confidentiality analysis, review and verification controls, vendor findings, a pilot, metrics, stop signals and a review cycle.
-
Choosing the Problem and Scoping the Plan
45 min
-
Controls, Vendors and Documentation
45 min
-
Implementation, Metrics, Failure Indicators and Improvement
50 min