With a busy start for my iOS programming work in 2019 I unfortunately hadn’t had time to organize events for Legal Hackers San Diego. However, the second half of the year should be busy and to kick it off we did a workshop on how technologies like machine learning and blockchain will affect legal operations and products. Here is the outline of the talk with links to key resources from that presentation on July 27th 2019.
- 10:00 am: Doors Open & Networking
- 10:15 am: Welcome & Introduction
- Introductions
- My Background: Notre Dame B.A. (Gov’t, Computer Applications, Japanese), Georgetown Law J.D., Cooley, iOS Programmer
- Legal Hackers: global movement of lawyers, policymakers, designers, technologists, and academics who explore and develop creative solutions to some of the most pressing issues at the intersection of law and technology.
- What is Legal Tech?
- Currently much of the tech involved in legal practice involves using MS Word and PDFs. Law moves slowly, is largely paper-based, and lawyers mostly charge by the hour so there is insufficient institutional incentive with law firms to innovate with technology to improve efficiency.
- However advances in technology are allowing clients to cut through some of this.
- Introduction of “Computational Law“
- Stanford Codex: Law School + Dept. of Computer Science
- “The branch of legal informatics concerned with automation and mechanization of legal analysis.”
- Examples
- Litigation: Predictive Coding (e-discovery)
- Corporate: Contract Analysis (Law Geex)
- Supreme Court Predictions & Stock (Dan Katz)
- Market Size
- Legal (US) $430B,
- 95% legal non-consumption? The idea that even if you have a valid legal claim, there may be as high as a 95% chance you don’t pursue it because of negative perceptions related to cost, complexity, and time required to do so.
- Introductions
- 10:25 am: Machine Learning 101
- Theory: Distinguishing Apples from Oranges
- The data problem. What rules would you write to tell a computer how to distinguish and apple from an orange?
- Maybe if image had more red pixels than orange, it’s and orange.
- Maybe if the item is cut open and it’s white it’s more likely an apple.
- What if the image is black and white?
- It’s hard to articulate the rules are brains use to do this.
- This is not how humans work.
- We have an abstract mental model of apples an oranges we can’t articulate, but have been developed from exposure to millions of known apples and oranges that allow us to classify a new object when presented with it.
- You can’t describe your model of an apple
- But this (image of square apple) would fit in apple even though it violates rules. We similarly train computers through machine learning by exposing them to known items, they develop a model not based on explicit rules we set and then can classify new items presented.
- Natural Language Processing
- Word Vectors: You can do math with words!
- The data problem. What rules would you write to tell a computer how to distinguish and apple from an orange?
- Practical Tools:
- What is Python?
- Anaconda
- Just like on a laptop where you start with Windows installed, you can add capablities to that device by installing additional software libraries. For example if you install MS Word, now you can do word processing. If you setup a printer, you can print, and so on.
- Python (and other languages) work a similar way. You start with the basic language capabilities but you can expand that capability with additional libraries. In our case we want to do natural language processing with Python, so we’ll add the Natural Language Toolkit (NLTK).
- (Lawyers are basically biological natural language processors)
- Jupiter Notebooks: This is a great tool for writing, executing, and sharing code for projects.
- Example
- Startup Examples
- Criminal Justice
- Educational Resources
- Theory: Distinguishing Apples from Oranges
- 11:00 am: Blockchain 101
- Theory:
- What Problem Does Blockchain Attempt to Solve?
- It’s a philosophy not a specific technology.
- How do Blockchains Work?
- Picture a Lending Circle – Susu for Criminals
- Centralized Model:
- Rules
- 5 people (all independent criminals) contributes $100 each month.
- Each month 1 person gets the whole pot.
- Rotate each month until each person has had a chance to receive the whole pot.
- A 6th person is put in charge of collecting money, keeping track of contributions, and making payouts (for a 10% cut).
- Problems:
- Everyone has to trust that 6th person to not run off with the money or otherwise manipulate the system.
- Everyone is losing 10% to this 6th person.
- Rules
- Decentralized Model:
- Replace the 6th person with a computer program (the “Smart Contract“) that runs on each persons laptop.Each laptop has a complete record of all contributions and payouts so to cheat the system, you’d have to hack multiple systems (not just corrupt 1 person).
- The program holds the money and makes the payouts automatically according to the rules established.
- Result: Group saves 10% and increases security.
- What Problem Does Blockchain Attempt to Solve?
- Practical Tools: Looking at an actual Ethereum smart contract.
- Atom (or other IDE): You could write the code in any text-editing application (Notepad, Word, etc.), but using a specialized Integrated Development Environment (IDE) gives you a lot of time-saving functionality.
- Truffle Framework: Some great tools for creating well-structured Smart Contract-based applications.
- Looking at a super-simple “Smart Contract”: CryptoSusu
- Integrating Legal Contracts & Smart Contracts
- Educational Resources:
- “Blockchain and the Law“: Great book explaining how blockchains work and how they will impact business structures and legal operations. Author is the founder of OpenLaw
- Stanford FutureLaw Annual Conference
- CryptoZombies: A fun way to learn how to write code creating smart contract based applications.
- PluralSight: Online course on building blockchain applications.
- My Fav. Podcasts on This
- Theory:
- 12:00 noon: What’s Next? Startup Weekend Legal Tech!
- Excited about the possibilities to create new startups in legal tech? Join us August 23rd – 25th at Walmart Labs for Startup Weekend: Legal Tech, Blockchain, and FinTech
- At Startup Weekend developers, designers, and entrepreneurs pitch ideas on Friday night, form teams, and work to launch new startups by Sunday. Along the way we plug them into experienced startup founders, investors, and coaches to help guide them to success.
- Here is a video from an earlier Blockchain Startup Weekend in NYC.
- Here is one from an event here in San Diego.
