Fear and Machine Learning: AI and Legal Practice

In late September leaders from legal, business and technology gathered in New Orleans at Clio Con, also know as the Clio Cloud Conference.* Andrew Arruda, ‎CEO and Co-Founder of ROSS Intelligence, delivered a talk called, “Artificial Intelligence and the Law: Science Fiction or Science Fact?

Arruda sets out to clarify some of the misconceptions that surround artificial intelligence and demonstrate how AI is currently being used. He provides examples of AI inside and outside of the legal environment and concludes that if it’s working elsewhere it can work for law too. He then talks about the fear piece and the “dangers of technophobia.” And he’s happy to report that he has so far been unable to find that “army of robot lawyers that are supposedly going to come and take our jobs.”

AI is a broad term, he explains, that encompasses the more specific process known as “machine learning” as well as advances in speech recognition, visual interpretation and natural language processing. With these technologies a computer interacts with its environment and human trainers and learns to answer questions and ultimately do things they were not initially “programed to do.”

Arruda begins his survey of AI examples by talking about the many “commonplace” self-driving cars he’s been exposed to since relocating from Toronto to Mountain View, California. Self-driving cars use a range of technologies (visual recognition, radar, sound, etc.) enhanced by machine learning which ensures that these cars “continuously get smarter and better.” As Arruda states with excitement, as each car learns from its mistakes it informs all of the other self-driving cars in the system so that they all “learn and scale out that learning.”

In the financial industry he compares high-speed AI traders to the experienced human trader who draws on a long career and a well developed sense of what to do when stocks trend in certain ways. However, “as humans we can only recall a certain amount of information.” AI systems on the other hand can not only “see thousands and hundreds of thousands examples of past trades” they can “instantly right collect them and then start to learn based off of the moves that they are making as well.”

In healthcare AI systems are used to help simulate drug tests or detect tumours in cancer diagnoses. For example, a doctor who may have limited resources and experience can submit an image of a patient’s tumour to an AI system which then compares it against millions of other tumour images and suggests possible outcomes and courses of action.

Arruda summarizes his talk to this point:

“So I looked over with you already what AI is, we chatted a bit about how AI is being used in transportation, in finance and in health care. Does any of what I’ve said sound outlandish and crazy? Like super sci-fi ish? Anything? This is all like, everyone’s like, that’s, that’s OK, right? Now, I’m gonna start talking about how AI can be used in law and let’s still remain, let’s have that attitude … come with me, because I set you up!”


This is what’s currently happening in law today:

  • ability to more efficient contract review
  • ability to make predictions on outcomes of patent cases
  • streamline due diligence efforts
  • streamline ediscovery efforts
  • improve and enhance legal research

This last one is the AI niche that ROSS Intelligence falls in to.

Arruda then highlights some recent reports that show a steady decline in growth in the legal industry and an increasing inability for firms to recoup the cost of the energy that’s necessary to prepare a legal argument or opinion. It’s 2016, firms can no longer afford to throw time and money at a problem, especially when in the end, they cannot even bill for it. The average associate, he says, is currently unable to bill for as much as $40K for legal research. For a 20 lawyer firm that equates to about a “million dollars of losses a year.”

Earlier in his presentation he asks: “Why do we sometimes think to ourselves: ‘Well there’s no way this technology can do anything in the legal space because what we do is so unique?’” But he counters with what he concludes should be the real question: “How can I start using artificial intelligence in my practice today?”

Ultimately this presentation is a pitch for ROSS Intelligence, a technology that “came out of the University of Toronto AI labs” and started as a dream “in a basement in Toronto.” Eleven months in they’ve moved to Mountain View, started selling ROSS and are contributing to changing the process of research in legal practice.

“[Technology is] there beside you,” he says, using KITT from Knight Rider as the analogy:

“… it enhances your abilities. It’s not replacing you. You still need that Hasselhoff mind to be able to do what you need to do … The human’s up front and the piece of technology is behind. It’s helping the human accomplish much more …”

And that’s nothing to be afraid of… right?


* Clio is a cloud-based law practice management platform launched in 2008. Their head office is in Vancouver and they have offices in Toronto and Dublin, Ireland.

Comments are closed.