Why we built this: Most AI training teaches lawyers the basics of how to use the tools, but we think it’s equally important to emphasize the role of judgment—not just the underlying legal judgment that must be developed alongside AI skills, but the judgment about when AI is the right call and when it isn't. Understanding that this would be difficult to teach through policy or lecture alone, we built an interactive simulation that puts judgment at the center and lays out the consequences of each choice.
AI × Law is a branching scenario simulation set in a fictional Southeast Asian regional law firm. It currently runs in two modules, The Associate and The Senior Associate, because the considerations around AI skill development, use, and risk change based on seniority.
The design rule throughout is that no choice is free. Every scenario offers four options, and each carries a genuine cost—professional, relational, or personal. Choices move visible trade-off meters so players can see what they optimized for and what they gave up getting it.
Module 1 — The Associate: This module explores the many factors junior lawyers will need to consider daily as they build the skills required to excel in the legal profession. Should you complete the mandatory AI training or meet an urgent deadline? When you cannot verify an entire AI-drafted memo, how much do you check, and how do you decide which parts matter most? A cited case turns out to be real but may not support the argument. Do you pull it, reframe it, or defer to the senior lawyer who already approved the draft?
Module 2 — The Senior Associate: This module explores how to manage teams that use AI while also building your own AI skills as a senior associate. Should you quickly fix a junior’s AI-generated error yourself, or send it back for correction and miss the deadline? A strong associate resists AI training. Do you mandate compliance, mentor her yourself, or work around her? You haven’t used the AI tool in months and now you’re reviewing work that’s generated by it. Do you delegate again, or close your own skill gap?
Built With
- Claude: Scenario design, branching narrative, and consequence writing developed iteratively
- Claude Design: Application design and the interactive build; no hand-written code
Time invested: We spent approximately 5 hours developing the application (design and game flow) and approximately 20 hours developing game content. Since this game will be used to train real associates at Tilleke & Gibbins, it was critical to get the content right.
Created by
- Andrew Stoutley, Chief Operating Officer, Tilleke & Gibbins (www.linkedin.com/in/andrew-stoutley/): Responsible for project concept and advisory, as well as incorporation into official training syllabus.
- Derik Schmidt, Manager of Strategy and Innovation, Tilleke & Gibbins (www.linkedin.com/in/derik-schmidt/): Responsible for product direction and execution, with a heavy emphasis on content design and development.
- Jidapa Eagark, Intern, Tilleke & Gibbins; LLM Candidate, University of Chicago Law School (www.linkedin.com/in/jidapa-eagark-29639a1a5/): Responsible for application and content design and development.