Earlier this year, Thomson Reuters announced a partnership with Watson, IBM’s artificial intelligence system. Tim Harty, Global Head of Actionable Intelligence, Thomson Reuters, discusses the future and present impact of AI on the law.
What can you tell us about Thomson Reuters’ plans in the cognitive computing space?
Thomson Reuters is committed to better serving its customers by providing cognitive solutions, and we are working on a number of initiatives in the legal and risk space. As part of our investment in artificial intelligence, Thomson Reuters is working with IBM Watson to accelerate the development of cognitive capabilities in our products. We have also recently established a Centre for Cognitive Computing which provides dedicated focus and resources to explore and develop in-house capabilities in the rapidly developing field of cognitive computing.
Video of our legal debate on AI’s impact on the legal profession:
When can we expect to see products in the market from this partnership, and can you give any clues about what shape they may take?
I can’t be too specific at the moment, but we are targeting to bring some capabilities into beta in Q4 2016, and launching a new product in the first half of 2017. We’re being ambitious in our plans, which means that to get it right we need to make a big investment in time and effort, in order that we can help solve problems for our customers. We are engaging with a number of customers around the globe to help design and fine-tune our offerings.
Why is it important for company like Thomson Reuters to be at the forefront of developments in AI and cognitive computing? Do you think it’s something that law firms and clients expect?
It’s important that we help our customers to solve problems and to succeed – if our customers succeed, we succeed. So, for example, two things our customers need help with are improving profitability and improving service to their customers. Cognitive computing can help in both these areas, for example, by reducing the amount of time it takes to complete certain tasks. Whether it’s the general counsel or compliance office servicing the business, or the law firm serving its client, we’re using AI to offer better answers more quickly, so that our customers win. That’s how we’ll win.
Many of our clients are looking at developing their own AI solutions with IBM and others. How might Thomson Reuters’ AI/cognitive project differ? What is Thomson Reuters bringing to the market that could benefit law firms and legal departments?
The investment required to tackle complex domains is significant, in terms of training the machine, applying the technology and designing user interfaces. This is an investment that TR has committed to making. Also, we can complement cognitive capabilities with our content and other value additions such as metadata, features and functionality. It’s the combination of content, metadata, features and cognitive that will help us to deliver significant value to our clients rather than cognitive on its own. Cognitive is not a product in its own right – it’s a capability.
AI is already here to a greater extent than people realise. For example, 15% of Google’s search queries are now handled by an AI (RankBrain). Has cognitive computing already changed the way Thomson Reuters delivers for its customers?
Thomson Reuters products have long included AI components. For example, in the Legal and Risk areas, in order to quickly find relevant documents, technology assisted review tags relevant documents based on a trained seed set. Also, legal research has been using sophisticated natural language processing techniques for many years.
Lawyers might be tired of hearing about AI at every event and conference. Why should they take notice of what we’re doing now?
We’re taking on some ambitious challenges and if we can solve them, our customers will be bowled over. There’s no guarantee we’ll succeed, but if we do, we’ll offer game-changing features in some existing and new products. That’s why our customers should be excited.
The AI/cognitive computing start-up landscape is already busy and exciting – what are the potential applications that excite you the most?
Two things really excite me. Firstly, the opportunity to help solve complex problems for our customers and to deliver them real value. Seeing the reaction of our clients to some of the tools and features we are building is energizing and exciting. Secondly, the prospect of being able to conduct non-menial tasks and jobs. Automating menial tasks is very well-established business practice: one of Apple’s main suppliers, Foxconn, has replaced 60,000 factory employees in China with robot workers. However, cognitive provides us with an opportunity to undertake non-routine tasks and jobs. Being a part of this, and witnessing to what degree this will be achieved, is fascinating.
In a recent podcast, Eric Laughlin (Corporate Segment Director and Legal Manages Services General Manager, Thomson Reuters, USA) mentioned that ‘showing our work is an important element to trust’. How do we make sure customers trust the accuracy and legitimacy of AI-driven solutions?
As much as we’d like it to be, AI isn’t magic. It takes an enormous amount of investment, effort and subject matter expertise to support the development of cyber physical systems. Thomson Reuters is fully committed to succeeding in offering cognitive computing capabilities that add significant value to customers, and part of that commitment is to make a significant investment in SMEs and in a Centre for Cognitive Computing to support the application of AI in our products. Product development involves extensive user experience design workshops and working with customer design partners to make sure we get the experience right. Products go through rigorous quality control and beta programmes to ensure they are market ready: we only succeed if we offer solutions that help our customers succeed.