Highlights
- Thomson Reuters chief legal officer, Norie Campbell, explains how general counsel can navigate AI amid widely varying knowledge and capability.
- Campbell's legal department runs Customer Zero Lab and serves as the field study for GCO 2030 research.
- Four principles give general counsel answers on data, security, trusted content, and verifiability when evaluating AI for high-stakes legal work.
I was already drawn to the role of Chief Legal Officer at Thomson Reuters. It was a dynamic and innovative company on the cutting edge of technology for the professions that it served, it had a compelling purpose that you could see its employees living every day, and I loved all the people I met. But what really sealed the deal was my inspirational CEO saying to me: “Norie, take this seat and be an ambassador to a changing profession.” What he could see, years ago, was how much change was coming to the legal profession as AI would make doing some of the most important work of lawyers quicker and better.
Since that day, I have been working on the question “What’s a General Counsel to think?” In my more than two decades as a general counsel, I have never seen a topic upon which there is such a wide dispersion of views, knowledge, and hands-on capability as the use of AI in the practice of law. And to me that makes sense—understanding AI feels to me like a full-time job. The good news for me, it is a big part of my full-time job at Thomson Reuters.
But if you are a bank general counsel, for example, you are running to keep up with changes in the banking industry and the law and regulations that apply. Where is the time to learn about a harness or a Model Contextt Protocol or all of the other seemingly inscrutable topics in AI development? The general counsel who get this right won’t just do their work, they’ll change what’s possible for their companies.
The right technology can change everything
The best don't just do their work. They change what is possible.
See how ↗
How our legal department is answering the question
So, in our legal department, we have been advancing on a few fronts in pursuit of this ambassadorial role.
- First, we are running Customer Zero Lab, where we are the first user of our legal products. We hold ourselves to the standard we ask our customers to trust, by using it first on our own hardest work.
- Second, we are the “field study” to the Thomson Reuters Institute work on GCO 2030—what would a legal department look like truly transformed by the use of AI and how do we get there?
- And, perhaps most importantly, to the question “What’s a General Counsel to think?” we are championing Fiduciary-Grade AI™.
To me, Fiduciary-Grade AI is the most important thing I can do to be an ambassador to my fellow general counsels. When I am in front of a group of GCs on the topic of AI, I see many people looking for help. Some of them tell me they feel like they don’t even know the questions they should ask, let alone the answers they should require. These are trained risk managers and changemakers in their own right, who really care about being right in the advice they give and who want to understand how infusing AI in their work will make that so. They don’t have the luxury of getting it wrong.
When we describe the AI we build at Thomson Reuters as Fiduciary-Grade AI, we are saying this is AI that is built for those who work to the standard of a fiduciary and who have important professional obligations.
Four principles that work as a cheat sheet
As I think about my ambassadorial question “What’s a General Counsel to think?” I see the four principles of Fiduciary-Grade AI as a kind of cheat sheet. It is designed for me if I were not the CLO of an AI company but instead was a CLO in banking, or oil and gas or pharma. To me it provides the answers to the questions I would be asking to make sure my data remained my own, my information was secure, my tool was working with the content I have always trusted in my legal practice like Westlaw.
And perhaps most importantly, it answers the question “Why would I stake my reputation on this tool?” I could verify it myself, understand how it came to the answer, and do a deep dive when I wanted to. I could also have confidence that around 1,500 attorney-editors at Thomson Reuters have been hard at work to make the tool the best it can be.
“What’s a General Counsel to think?” In this unique time in the history of technology, where the best tools in the world are being purpose-built for lawyers, the general counsel needs to think about Fiduciary-Grade AI.
