The Tubman Project
So, I guess it’s time I write about the Tubman Project…
For the past year, I’ve been working on an initiative to correct the criminal justice system. For those unschooled in the works of Michelle Alexandar and James Foreman, we’ve got a bit of an issue… Our system is not only wholly inefficient, but it’s also damaging us in real and visible ways.
I can’t do much the times that it outright kills innocent people. I’m a technologist, not a superhero. But felt like there was something I could do about the other things.
Law was the obvious area to work. Being a coder, that’s just my language. Rules… Systems… Inputs… Outputs… It’s all there. And it’s an area that’s ripe for technological disruption. If you’re a regular reader, you’ve seen my many forays into what blockchain technology can do to voting and record keeping. One of the other major trends sweeping the computer technology world is AI. And it’s a tech that’s also making waves in legal tech.
AI Judgement software is already making its way into courtrooms in Cleaveland… Between them and the legal data companies, our system will change. The question is, will the human element keep up?
Enter the Tubman Project. The goal is simple: reduce the cost of legal defense. By creating open-source legal technologies that leverage data machine learning, we can prevent income from becoming a barrier to justice. And at the same time, we can democratize many of the technologies that are changing our world in a real way.
The kickoff for this effort was our first Hackathon in Houston. In the aftermath of hurricane Harvey, civic-minded Houstonian developers, designers, and lawyers got together to take the first crack at hacking the Justice System. We spent a month brainstorming and pulling in the talents of Civic Hacking groups like Sketch City and the Texas Criminal Justice Coalition to examine how to tackle the problem. Then, we set that creative talent free one Saturday at Station Houston.
This year we took what began in Houston and started to expand the community. I traveled to some of the Legal and Technology innovation hubs around the country to start connecting the pieces. I liaised with groups like Measures for Justice, CALI (Computer-Assisted Legal Instruction), MIT Media Lab, and Virginia Court Data. I spoke with prosecutors and public defenders. I began to build a picture of what Tubman needed to be.
And in the meantime, we continued to work on the technologies created at the first Houston hackathon. The fun moved to Austin where we participated in their yearly Hack for Change event.
All of that culminated in our most recent major hackathon. With the help of David Colarusso of Suffolk LIT lab, I narrowed the field to three major areas where we needed to make progress. Video Processing, Legal Knowledgebase, and Secure Communications. With those areas in mind, launched our second hackathon in Boston Massachusetts. I and our sponsor, Owlfa, staked $3000 against our three objectives. The team that was able to create the best solution to our three challenges would take home a $1000. And so, we recruited machine learning engineers and legal innovators from across the city to help move the chains.
The Competitors came from all over Boston and Cambridge. Some were students working on Law degrees or exploring the fields of computer science. Some were Lawyers looking to give back. Some were professional engineers working in the vast array of innovative companies in the Bay State. But all did their best complete the challenges and out-code their peers. In the end, there were two teams standing.
Metropolis
The first winning team was Metropolis. Varun Mazumdar, Matt Dotty, and Russ Mattson. This team built a machine learning algorithm that is able to pick out an identify pertinent areas of long steady video footage.
Their work leveraged some of the existing open source tech that has been created and applied it to our challenge. Using code from Open CV and other open source packages and data sets, they were able to build and demonstrate a Video Marking command line tool. This is a little program that can be executed by using your Windows or Linux based command line. (Try Windows Button + R and then type “cmd”. Have fun!)
During their demo, they showed off their program by picking out the important parts of a video in real time. Their program not only marked the frames that were important to watch, but it also was able to identify objects in the scene such as cars or people. You can view the demo here.
Alibi Checker
The second successful team was made up of a single Law School student, Niharika Singh. She worked on the problem of using the data cloud that’s created by big tech companies to your advantage.
Her project focused on Google. You see, Google Maps gives you the ability to record all of your location data and store it in Google Maps. It’s opt-in and deactivated by default, but if you choose to activate it you’ll have a running history of where you have gone and how you got there. While not perfect, I’ve been able to use the service and have it detect a number of modes of conveyance, from driving to taking a train… even riding a bike. After a while, you have a pretty comprehensive record of your travels.
While this may be more than most people want to record online, it does present an opportunity if you need an alibi. The question is, how do you leverage it in a way that keeps you safe?
Niha’s solution was to apply cryptography to share a limited bit of her location data when needed. Not enough to pinpoint her, but just enough to disprove an opposing theory. In this way, she can present a provable and third-party validated alibi even without a human witness by leveraging her phone. To see this in action, check out the video here.
The Survey Bot
The final prize proved to be too much for the groups that attempted it. The goal was to have something that incorporated an expert system in the form of a quiz or survey that could be taken online. While we had a number of contestants with technical backgrounds, very few had the legal chops needed to build a good Survey bot. Those that did (such as Russ Mattson) were hard at work on some of the other projects. So for now, it remains unclaimed.
But this is only the beginning of our work in Boston. Next year, we’ll be working with Suffolk Law School again to launch the Legal Innovation and Technology Conference in April of 2019. We’ll be holding workshops for legal technology to help attendees learn skills that they can apply directly to their work. We’ll also be hosting a hackathon where the finalists will present their projects during the first night of the conference on April 8th.
Until that time, watch this spot. And take a look at some of the other areas that the Tubman Project lives. Sample some of the podcasts that discuss it. Who knows… One day, it just might have your back.
Social Media Profiles