Many years have passed since Sir Tim Berners-Lee used to index the Internet by hand. Since then, search engines became our main tool for finding the stuff we need in this endless ocean of data. We all got so used to it, that we hardly ever think of how incredible it is that a query we search using Google, for instance, generates millions of search results in under a second. let’s face it, products that just search have lost their wow factor along time ago – we are now in the age of AI.
Personally, after years of thinking AI is just an excuse for creating lousy movies like I Robot, I finally figured that there is really something to it when I saw IBM’s Watson beat two of the most successful Jeopardy! contestants in 2011. This was the first time I realised machines can out-smart humans, and can actually understand rather than just read. 5 years have passed since, and the technology just keeps getting better.
From the basement to Silicon Valley
Influenced by this amazing progress in the field of AI, a computer programmer and a young lawyer from Canada have partnered to found the first ever legal tech company. “I had a really good friend that I met at the University of Toronto. He was a computer scientist, particularly interested in AI”, told us Andrew Arruda, ROSS’s CEO the other day. “At that time I was working in a small law firm in my neighbourhood for about 6 or 7 years, and this friend called me and said ‘hey, what do you think about a system that will allow lawyers to ask questions in the law in natural language and get their answers instantly?’. I said that this is something that is definitely going to happen, and let’s be the people to do it”.
From that point the story of ROSS rolls out exactly like the classic basement-to-Silicon-Valley startup story. “We started off working from the basement, maxing our credit cards just like you hear in the stories, until eventually we got our first funding from Y Combinator, packed up our bags and went to Sillicon Valley”. Besides the buzz the young entrepreneurs have managed to create in the tech scene, another big push came from an unexpected direction. “At that point we raised another investment from Dentons, the world’s largest law firm”, Andrew says, “that basically showed us that not only the tech world was excited about what we were doing – the legal world was excited as well. Having the financial backing of the world’s largest law firm really showed us at that early stage that we were on to something”.
Siri who graduated from Law School
The system ROSS developed takes advantage of the huge technological progress made in the field of AI in recent years, and allows lawyers to ask questions in natural language and get short, accurate answers with the sources to go along with it. The database behind ROSS is consisted of court rulings, articles and other legal sources, and in contrast to big known legal databases such as LexisNexis, ROSS is not about searching data – it’s about brining insights. “Unlike these data providers, ROSS doesn’t provide lawyers with huge amounts of raw data and leaves them to deal with it by themselves, rather it provides them with insights. The system knows how to “read” content quickly, extract the relevant passages and comply answers in reasonable language. That allows its users to work much more quicker and much more efficiently. The other key difference is that ROSS can also get smarter. by interacting with the system and up voting or down voting the answers they receive, the users actually teach ROSS, thus making it smarter every time they use it”.
And indeed, ROSS’s learning abilities is probably its most promising feature. Imagine a scenario where a group of lawyers are working with the system over a period of time, in which they constantly rate its answers’ relevancy, based on their precision, credibility and focus. The system’s technology allows it to imitate human learning abilities – it learns from which databases to “pull” which content, how to “understand” different questions, how to interpret various terms and so forth. One can only imagine the level of efficiency this type of technology can achieve combined with constant human feedback.
First client the past month
During last May, the company announced it has been hired by Baker&Hostetler (an American law firm ranked in 2009 as one of the 100 largest firms in the world) for its bankruptcy department. “This is a big deal for us. We have several other firms and we’ll make some more announcements shortly”, Andrew says. “For us what it really shows is that the market is ready for a solution like ours, which really is about driving efficiency and allowing lawyers to do a lot more with less. Right now we’re in a really competitive landscape, I’m sure you see it in the Israeli market as well, there’s a lot more pressure than it used to be, and so law firms are now really more open to these new technologies to help them stay competitive”.
It has been argued numerous times in the past that law firms tend to be really conservative and slow in adopting new solutions, certainly ground breaking ones like ROSS’s. “There’s only so much one human can do, and with that it drives up the cost of legal services. That means that a lot of people who need a lawyer can’t afford it. Using artificial intelligence products like ROSS simply means that law firms can deliver the same great quality in less time and less money. hopefully that means that more people who actually need a lawyer can afford one”.
AI in Hebrew – around the corner or a far dream?
Being a technology that knows how to “read” and “understand” text, language is a crucial aspect in AI. While there is obviously a financial logic in investing on a system for 1.5 billion English speakers world wide, it seems that making the same investment in a system aimed for no more than 8 million Hebrew speakers around the world is much less reasonable. “There is a commitment from all the companies providing AI software, such as IBM and Watson, Facebook, Amazon, Google and others, to include as many languages as possible, so I think it’s just a matter of time until all languages will be served. With that being said, I think it can take a while, especially for uncommon languages such as Hebrew”, Andrew explains. And indeed, a quick look at Watson’s supported languages shows that as of today the support for different languages is very partial.
However, there is no reason for being pessimistic. The adjustments required to make AI softwares work with other languages does not mean starting from scratch. The constant improvement in AI availability certainly brings this day closer.
The vision – creating more tools for lawyers
What can we expect seeing from ROSS’s direction in the near future? mostly more tools, more areas of practice and a lot more lawyers using it to improve the way they do their job, Andrew says. “Our vision is to expand to as much areas we possibly can, and offer as many tools possible to allow lawyers around the globe to do more in less. We have some fascinating products we’re cooking this days, and you can expect hearing about them and about other law firms that have decided to work with us real soon”.
The year of 2016 was forecasted by experts to be the “year of AI”, and in the meanwhile it sure looks like a good forecast. It is true though, that we’re not expected to see any similar products in Hebrew in the near future, but is seems as if Legal Tech enthusiasts around the world have a lot to expect.
The title picture is courtesy of O Palsson.