Q&A with David Wong, Thomson Reuters Chief Product Officer
In the span of roughly six months, generative AI – a form of artificial intelligence that can compose original text, images, computer code and other content – has gone from tech novelty to the biggest business disruptor since the advent of the Internet.
After handily passing the bar exam, acing the SAT, and convincing the world that Pope Francis had suddenly developed a penchant for Balenciaga jackets, many commentators have started to wonder what the technology will mean for the future of work in professional services. Will lawyers, accountants, computer programmers, journalists and countless others suddenly find themselves obsolete?
David Wong says generative AI will not replace these jobs, but it will fundamentally change the way many of them are done. He should know. Wong is the chief product officer at Thomson Reuters, a company that has been building AI-enabled software for decades. It just announced a commitment to invest about $100-million a year to incorporate generative AI into a wide range of products, including the leading software solutions used by legal and tax professionals to research case law, draft contracts and manage tax compliance globally.
We sat down with Wong to talk about how generative AI is being used today and how it will affect professional services jobs in the near future.
Q: The hype around generative AI solutions like ChatGPT, Bard and Bing Chat has been impossible to avoid. Is this technology as disruptive as everyone says?
A: Generative AI is definitely a disruptive technology, but not in the way that many people are hyping it up to be in some of the more sensational stories about bots putting highly trained professionals out of work. In the near term, the biggest disruption we’re going to see is to businesses that build services or products that involve retrieving or presenting information or helping to author content or work products. Anywhere you currently see a search box or someone extracting and producing summary content—whether that’s audio, video, the written word or code—you can expect day-to-day workflows to change pretty significantly. That doesn’t mean removing humans from the loop, though. It means the way they do their jobs will change.
Q: How are you implementing generative AI in the products you develop at Thomson Reuters?
A: At Thomson Reuters, we’ve been leveraging AI and machine learning in our professional information products for decades and we first started experimenting with incorporating generative AI into our solutions back in 2020, when the technology was still very new. That head-start has really positioned us well to start integrating generative AI into several of our research and drafting solutions.
Q: What are some of the biggest changes that users of your products should expect? How will the technology impact their day-to-day lives?
A: Our entire approach to generative AI starts with our customers. We’re looking at their biggest pain points and designing solutions to help them spend less time doing repetitive, labor-intensive tasks while improving the quality of their output. By focusing our strongest assets—our proprietary research databases and our human expertise for training and tuning these generative AI models—on these challenges, we will dramatically improve our clients’ day-to-day workflows in a few important ways. First, we will create time savings and efficiency by getting clients to the right answer faster. On top of that, we will help our clients reduce errors and dramatically decrease the likelihood of missing something in their research, discovery or auditing processes. Perhaps most exciting of all, though, we will help to improve collaboration and introduce entirely new ways of helping clients expand their aperture for creative problem solving. This is about much more than just slapping an AI-powered search box on our products. We see this as a critical step forward to creating fundamentally new ways of working together, sharing information and tackling complex challenges.
Q: What about some of the problems we’ve heard about with generative AI, such as examples where the bot confidently answers a question incorrectly? What are you doing to address those issues?
A: Access to reliable, authoritative reference data and training are the two keys here. In practical terms, that means generative AI use cases focused on a highly technical, specialized subset of data like case law must be able to synthesize existing large language model data with proprietary content databases to deliver concrete citations to support their outputs. Importantly, that process also needs to be intermediated by human subject matter experts who understand the nuances and the context to train and fine tune the results. That’s precisely why we’re so excited about the prospects for generative AI for legal and tax professionals because we are uniquely positioned to deliver both. Our legal and tax and accounting research databases are the gold standard and our huge teams of legal and tax professionals hold the keys to fine tuning these algorithms to be incredibly powerful and accurate.
Q: You’ve been working in technology for your entire career, with stints at McKinsey, Nielsen and Facebook, among others. How significant a development would you say generative AI is relative to, say, social media or digital video or some other transformative technologies?
A: The closest parallel between what’s happening right now in generative AI and previous tech disruptions, in my opinion, is the advent of the smartphone. In both cases, we’re dealing with a technology breakthrough that enables entirely new categories of solutions that would never have been possible without the base innovation. Applications like Uber and Instagram, for example, could never have worked without the smartphone. Right now, with generative AI, we’re in a phase similar to what we saw in the early 2010s when everyone was scrambling to launch an app. In this case, everyone is scrambling to create an AI-powered chat function. The real breakthroughs will come when generative AI enables new types of solutions that haven’t even been imagined yet.
Q: What aspect of the ongoing evolution of generative AI has you excited about the future, and why?
A: I am most excited about the critical role my colleagues and I will get to play at this pivotal moment in history. Thomson Reuters is in a really unique position to help our customers not just change but transform the way they work while improving their lives along the way. And we have the assets we need to build those solutions right now. This really is an incredibly fun time to invent new products and I feel lucky to be a part of it.
Advertising feature provided by Thomson Reuters. The Globe and Mail’s editorial department was not involved.