Global AI Regulation Trends
In this week’s episode, Katherine and Anna discuss the differences between AI regulatory regimes around the world.
- Guests & Resources
- Transcript
Partner
» BiographyCounsel
» BiographyKatherine Forrest: Hey, good morning, everyone. We're here today for another episode of “Waking Up With AI,” a Paul, Weiss podcast. I'm Katherine Forrest…
Anna Gressel: and I'm Anna Gressel.
Katherine Forrest: And we're your hosts for “Waking Up With AI” today. So, Anna, today is a day that I think we've all been waiting for, where we're diving into the topic that you and I talk an awful lot about, which is the worldwide regulatory environment. And I've got two questions for you.
Anna Gressel: Oh, you do?
Katherine Forrest: I do. I do. The first one is can we do this in five to seven minutes? And the second is, is this going to be deadly boring?
Anna Gressel: I believe we can, and it'll definitely be fun. And we make AI fun, we have fun every day, Katherine, doing this together.
Katherine Forrest: All right, all right, let’s get ready to go and let's start with an overview of where things are headed in the world today. And right now, we're in the second quarter of 2024.
Anna Gressel: We're seeing a ton of regulatory activity from all over the world. I can tell you this moment is different from before. There are proposals every day, but let's talk about a few things. There's the EU AI Act and the White House Executive Order. We'll talk more about those in a minute.
Katherine Forrest: And there's also activity in a bunch of other areas. I've recently been talking about activity in Japan, the UAE, Saudi Arabia, Singapore, South Korea. And Anna, when you look around the world, can you find any general themes that are shared in the different regulatory environments?
Anna Gressel: I think we tend to see them bucket into two categories. We have our softer touch and lighter touch jurisdictions. That might be like the UAE and Saudi Arabia for right now. Singapore too, they have voluntary guidelines, and they're really focused on innovation and investment in technical advances and responsible innovation. They do include voluntary guidelines.
Katherine Forrest: There's also sort of another side to that, and that's the parts of the world that have a more formalized regulatory regime, and I'm thinking of the EU AI Act being primary among that group.
Anna Gressel: Exactly. Where would you put the White House Executive Order?
Katherine Forrest: Well, I would put the White House, they call it the White House E.O. for Executive Order or the Biden Executive Order sometimes. I'd put that closer to the EU AI Act, but without the formalization or the attempts to really spell out compliance method in a formalized way. The Executive Order is basically a directive by the White House for different branches in the executive branch overall to do a variety of research as to what kind of compliance is going to work, to assess things, to bring stakeholders in, and to then go ahead and implement. So, privacy, security, algorithmic bias, those are themes that are flowing throughout the Executive Order.
Anna Gressel: And let's talk a bit about the EU AI Act. Katherine, you and I have been doing this for a long time. You and I have been working in this space since 2018 so I think we both remember when the EU AI Act’s white paper was first introduced back in 2020. We've been hearing about it for years, but the technology has changed a lot in that time. So, it's also been adjusted in real time to accommodate that.
Katherine Forrest: It really has. And I remember speaking at a conference with you in Brussels where we were talking about narrow AI. It was before generative AI had even hit the scene, and we were talking about how even narrow AI was going to effectively be regulated. But now we've got generative AI and the EU AI Act has expanded to encompass that. And in March and April of this year, it hit a really important milestone.
Anna Gressel: Yeah, I mean, now it has been passed and it takes effect over a period of time. Some of those deadlines are going to start pretty soon.
Katherine Forrest: Right, I think some of the provisions, well I know some of the provisions actually start as soon as six months from now and others phase in over a period of up to almost three years.
Anna Gressel: Katherine, what are some of the provisions that phase in right away?
Katherine Forrest: Well, the provisions that are going to phase in right away are mostly the prohibitory provisions. The act is based on assessing use cases by risk level and some of those AI use cases are on a prohibited list.
Anna Gressel: That includes AI for police surveillance and social scoring. Let’s talk about how enforcement is going to work. Each country is tasked with developing its own local enforcement office. That means it has to fund that office and have the skill set to monitor the conduct of AI usage in its area. Katherine, does anything concern you about it?
Katherine Forrest: I wonder about a couple of things. First, as we just talked about a little bit earlier in this episode, the technology is changing all the time and we'd mentioned when we were in Brussels talking about the EU AI Act when really only narrow AI was at issue. And now we've got generative AI and we don't know where AI is going to be in 36 months or some period of time as this EU AI Act phases in. So, I really wonder how the regulations as they exist now are going to map onto new technology that we don't even understand. And I also worry a little bit about countries that are tasked with monitoring and forcing AI. Having the resources to be able to hire a sufficient number of technically proficient personnel, who are going to be able to do the work to achieve the compliance that the EU AI Act sets up.
Anna Gressel: And the EU AI Act also extends to companies doing business in Europe, as well as companies doing business outside of Europe. It's kind of like the GDPR in that way. So, for folks in the audience, our practical pointer is to start doing a deep dive into these regulations and assessing how you can conform your existing compliance programs to meet these new EU AI Act requirements and some of the requirements that may be coming down the road from other jurisdictions.
Katherine Forrest: There are a number of businesses that are doing that right now, and there are practical things that can be done to get companies ready so that they're well positioned for the go live date. What's the first thing that should be done from your perspective, Anna?
Anna Gressel: I think first is assessing your use cases and developing a matrix of where they fall within the EU AI Act's risk assessment scheme. That'll tell you what you're looking at and what the timelines are that are actually going to apply to you as a company. I think that's all we have time for today, Katherine.
Katherine Forrest: It is, and I want to remind the audience that we're not giving legal advice. Anna and I are just here talking about the EU AI Act and the regulatory schemes that are across the world, and we encourage you to take a look at them yourself and form your own views. I'm Katherine Forrest…
Anna Gressel: I'm Anna Gressel, and we hope you enjoyed our most recent episode of “Waking Up With AI,” a Paul, Weiss podcast.
Katherine Forrest: Alright folks, we're done for today and we hope that you subscribe to the podcast. We'd love to have you as subscribers and you can find us on Spotify, Apple, Overcast, and the Paul, Weiss website. Have a great day.