Remy Blaire: Well, corporate America is under pressure to prove a return on its heavy AI investments, but corporate governance isn’t succeeding in keeping up. Well, four in five organizations are deploying gen AI, but only 30% have comprehensive policies on how it’s used. More than 80% of public company directors admit to using AI tools for board materials, yet barely 6% of corporations have formal rules governing director use. And this is creating severe risks around trade secrets and attorney-client privilege.
Well, joining us this afternoon to break down boardroom blind spots as well as enterprise risk is Kaitlin Betancourt, who is a partner at Goodwin and also co-chair of the firm’s AI practice and chair of its AI governance committee. Kaitlin, good afternoon. Thank you so much for joining us. So for the companies that deploy first and write the policy later, where does it actually go wrong?
Kaitlin Betancourt: Thank you so much for having me on, Remy. It’s a pleasure to be here. Yes, that’s right. So what we’re seeing is that there was mass adoption of AI, and there continues to be. And that’s sort of leaving legal departments and risk and safety departments with a lot of components that they need to get their arms around, in order to make sure that companies are staying onside, you know, in accordance with the commitments they’ve made to customers, and monitoring risk in a way that is really appropriate and defensible.
So we’ve identified a number of spots for improvement for companies, at the company level and at the board level. And it’s really a multilayered approach that we’re advising on. So there’s plenty in the AI governance space now for companies to stay abreast of. And what we’re seeing is that governance is really what protects the company’s return. It helps the company move fast, rather than, you know, obstruct the adoption of AI. So it’s super important for responsible AI adoption, but it also helps the bottom line.
Remy Blaire: Yeah. And I do want to zoom in on something you just mentioned, and that is the word risk, in particular legal risk. So given the fact that AI agents don’t just answer questions, they also take action, how does that change the legal risk for an organization?
Kaitlin Betancourt: Yeah. So agentic AI really puts a spotlight on human-in-the-loop, ensuring that there are proper controls in place, proper governance controls and proper technical controls, to really, you know, ensure that agents don’t go off the rails. And, you know, permissioning is important. Access controls are important. But agentic AI highlights the fact that, you know, AI governance is so critical to being able to employ AI in a way that doesn’t put the company in a significant risk posture.
Remy Blaire: And Kaitlin, we know that people use artificial intelligence for all sorts of tasks. But when it comes to directors using AI for board work, apparently, if this is correct, only 6% of companies have actual rules for this. So are board members out there inadvertently waiving attorney-client privilege, and could they be leaking some sensitive company data?
Kaitlin Betancourt: It’s definitely an area we’re focusing on, ensuring that board members have the education and guidance they need to exercise their fiduciary duties in terms of their usage of AI in connection with company materials. So, you know, there are two components to AI governance at the board level. One is the board’s actual oversight of the company’s usage of AI, because as AI permeates company operations, that becomes a more material issue that warrants oversight by the board. And then the second component, which is what you just highlighted, is that board directors may be using AI and need to be privy to, you know, what is really permissible use of AI with the company’s sensitive information versus what isn’t. So we are advising boards on this very issue.
Remy Blaire: Yeah. And Kaitlin, here on Wall Street, we’re paying attention to artificial intelligence, the technology, the investment, as well as the trade. And there’s one company with a highly anticipated IPO, and that is Anthropic. I understand that Goodwin’s AI governance program was recently adopted by Anthropic. So what does an AI developer need from a governance framework that a typical company does not have?
Kaitlin Betancourt: Thanks for highlighting that. We’re so excited. Yesterday, as you mentioned, we launched Goodwin’s AI governance benchmark inside Anthropic’s Claude for Legal. And so, in about five minutes, it shows a company where its governance gaps are and what to fix first, which is really critical. So AI governance is iterative, and it needs to be, you know, looking at what are the weakest links, how can we fix those, and how can we move on? And so this is what our governance benchmark provides for companies in just a short amount of time. It’s clear what makes sense to focus on first, so that properly addressing AI governance does not become an overwhelming and impossible task.
Remy Blaire: Well, Kaitlin, we will have to leave it there for today. But thank you so much for joining us on this midday show today, and I appreciate your time as well as all of your insights.
Kaitlin Betancourt: Thank you very much. I appreciate it.
