← The Digital Executive
The Digital Executive
Alfredo González Briseño on The Missing Ingredient in AI Regulation | Ep 1292
00:29:01
Loading summary
Transcript14 lines
- [00:01]
A
This episode is brought to you by Accenture. When your advertising operations fall out of sync, everything else follows. Spotify and Accenture are working together to reinvent the rhythm of ad sales using automation, analytics and smarter workflows to simplify campaign delivery and access better data across the business. The result? Less time spent on operations, more time connecting brands with the moments and fandoms that matter most. Learn more@accenture.com Spotify
- [00:31]
B
this episode is brought to you by Google Chrome. You think you know a browser, but Gemini and Chrome? That's new. It can help you with practically anything on the web, like restoring a vintage motorcycle from a 50 page restoration block, or finally break down that long article you've had open for weeks. Gemini and Chrome is here for it, ready to make anything online make sense. There's no place like Chrome. Check responses, setup required compatibility and availability various 18 foreign. Welcome to Coruscant Technologies, home of the Digital Executive Podcast. Do you work in emerging tech? Working on something innovative? Maybe an entrepreneur? Apply to be a guest at www.corazon.com brand welcome to the Digital Executive. Today's guest is Alfredo Gonzalez Brinzenio. Alfredo Gonzalez Brinzenio is a policy innovator and expert specializing in the governance of regulation, including regulation of artificial intelligence. He has advised governments worldwide on business environment reforms, focusing on the how to regulate designing more efficient, transparent and predictable regulatory frameworks that build trust. He is the author of the recently published book Better Regulations, Better AI Results. The overlooked policy framework governments and policymakers need to regulate an AI driven world. Alfredo is also the founder and CEO of Fairly Simple, a govtech startup in Virginia. Using AI to transform complex regulations into something simple, clear and accessible so people can make better decisions. Well, good afternoon Alfredo. Welcome to the show.
- [02:12]
A
Hi, good afternoon Brian. It's a pleasure to be here with you today.
- [02:17]
B
Absolutely my friend. I appreciate it. And I know you're in Fairfax, Virginia, just an hour time difference from Kansas City. I appreciate that. I know sometimes it's hard to get on calendars and sync things up, so thank you again Alfredo. Let's get into your first question here. You've advised governments around the world on regulation and policy reform. What led you to this career path and the work you do today?
- [02:43]
A
Yeah, thank you again Brian. Well, sometimes our path started before we even realized they had. And while my former work on improving regulations started in Washington, D.C. when I was with the World bank there, I advise governments worldwide on how to improve their business and regulatory environments. Not one regulation at a time, but just thinking about governance Systems that produce better regulations in general. And at that time is when my industrial engineering and public policy backgrounds are connected and have been working together since then. But let me be a little bit more specific about what I mean by improving regulations in a systematic way. So for starters, you don't write a piece of regulation, sign it and hope for the best. And let me do an analogy with what a business project requires. So I know many of you in your audience are in the private sector. So as with any business initiative, you go through different processes to create something you need to proactively manage. Also the life cycle of this initiative, since it's an idea, you understand the problem you aim to solve. You try to understand how big is the pain your potential customer is going through and whether it's worth solving that problem or there are other solutions in the market. And when you decide to go for it, then you incorporate the feedback you have received from others from your customer discovery conversations to build and launch a product that has a market fit. And once it's in the market, you monitor for performance, look for improvements, make updates when necessary. So this same process can be applied for regulations and policies. So you follow a process that involves several stems. This is a good process for achieving good regulation. So it starts with the ideation, okay, we want to create this rule. It starts with planning, announcing what you're going to do, understanding the policy problem at stake, and what are the alternatives to solve that problem. And you try to use evidence while you do it. You do consultations within government, with other stakeholders, within industry. And when you make a decision, you stick to principles of transparency and full disclosure, so everyone knows what's happening and when the regulation is being implemented. You aim for simplification and at some point you need to go back and review to understand whether it still has a market fit. So again, this description, just as a business initiative, this implies that you need to manage actively a regulatory life cycle, not only for a few regulations, but for all of them. Because if your purpose is to creating a better business environment, you need to do this. And it's a simple process, but I have to say, it's not always easy. So having said all of these, how this path led me left to my work today. So to be honest, at some point the engineering side me kept calling and asking me to still solve problems like I was doing in college, and create new things. So I decided to do this with the knowledge and experience I got in Washington D.C. and international development, or how regulation usually works inside governments around the world. So today My path continues as a strategic policy advisor, as an entrepreneur, an author. And that's why in the past year I wrote Better Regulations, Better Results, launched my govtech startup Fairly simple, and have been creating tools and frameworks for leaders, teams and the public to pause first, clear out the noise of what's being said in the media and in different conversations, to finally understand regulatory decisions as they are and once you get there for them to plan for better AI governance and policies in their organizations.
- [06:27]
B
Thank you, really appreciate that. And I appreciate your backstory. Obviously from engineering, working in regulatory environments in the financial industry in D.C. you talked about your career there, but you correlated. You kind of. There's this similarity in the business world. You talked about process and regulation and policy. It's very similar to business. And of course there's very specific way to manage a regulatory lifecycle as you talked about. And of course you now are helping because you've created this policy framework for governments and policy that you're helping governments around the world to again ensure that there's better compliance governance around these regulations. As you know, it's very strict in the government sector and the financial sectors as well. So thank you Alfredo. Why do you believe the how of regulation is often more important than the regulation itself?
- [07:19]
A
Thanks Brian. This is an important question not only for AA regulation but for regulation in general. And let me make an analogy with high performing quarterbacks. So you are in Kansas and you see my homes every Sunday. My Raiders now have Fernando Mendoza. And what I'm trying to say is that high performing quarterbacks or athletes in general, they just don't show up 30 minutes before a game and just throw passes. So they follow a process. So during the week they have to recover from the game. They need to get ready physically and mentally for Sunday. And even on game days they follow a process like when they get on the bus, when they arrive to the stadium, even when they jump on the field and before taking every single snap during the game. So the same things happen with regulation. So you don't grab a piece of paper, draft your regulation and approve it immediately. Good regulation requires a process tool. So as I alluded a little bit before, it starts with planning and announcing future regulatory changes. Canada is a very good example of how to do forward regulatory planning. Then before you start drafting your rules, you need to understand the underlying process and even question if your regulation is the best solution in the process. You consult within the government, you consult with industry and other stakeholders. And this is very important not to get consensus, but to listen and make better decisions, and then you communicate final decisions in a way that is transparent. And when you implement them, you aim for simplification. And when it's possible, you provide flexible frameworks for compliance. And finally, you need to set a timeframe to review whether your rule is still relevant or needs to be updated. And this is very important for technology and AI because it's changing so fast that you need to get in the rhythm of reviewing rules, like sometimes every 6, 12 or 18 months. Again, despite being simple, this agenda is not easy to implement. It requires order, coordination, dialogue and communication, and to be honest, a lot of energy. That is why it's also important for governments for countries to have institutions leading this better regulation agenda. You cannot expect to have good regulations and just for them to happen spontaneously. So let me bring down these ideas to concrete examples we see today. I know you're in government, and at some point I'm curious to hear what you see there. From my experience, some governments embrace this agenda, that it's simple and they realize that it's difficult implementing it, but they still do it. Others just drop the ball because that's the easiest thing to do. And that's very common in many governments around the world. However, I know your audience is mainly in the US And I want to reflect briefly on a couple of recent examples here to share what happens when there is not a solid how behind regulatory decisions. So I'm going to use two cases, the recent case of Anthropic and the Export Control Directive and the recently approved AI Safety law in Illinois. So first, I want to acknowledge that these decisions and legislations may be responding to very valid and good reasons. We're talking about national security, people's safety. So again, that's not in question. I will be just making a reflection and analysis of how those decisions and regulations were made. So let's start with the case of Anthropic. So the decision was as unpredictable as it can get. So first, the actual text of the Export Control Directive and the decisions to leave the licensing requirements on Mythos 5 and Paywall 5 a few weeks after were never publicly disclosed. And even when the alleged reason behind these decisions was national security, the actual problem was never clearly explained and with strong evidence. At least that's what we know from Tropic's announcement on June12. Also, when licensing decisions are announced or any other regulatory decisions are announced on a late afternoon on Friday, you clearly see a red flag. In other countries where I've had conversations with the private sector, they call these type of decisions, weekend rules, and obviously they take everyone by surprise. Finally, when decisions with significant consequences are reverted a few days or weeks after, it tells you that most likely they were not well planned and the alternatives were probably not well assessed. So in the coming days, I will be publishing a more detailed analysis on this case of anthropic. So the second case is the AI Safety Measures act from elinois, approved just a few days ago. And in this case I want to run a different type of analysis based on the actual text approved. So if you go online and you read the law, it's 50 plus pages, you start seeing a few things. And this is what I see. So first, you will not find a single article or paragraph explaining what the purpose of the law is and the problem it's trying to solve. Of course, that has been communicated in media that can be infirm. But it's always a good practice to have in those documents, in those legal and regulatory documents, a clear statement of the problem in question and the policy objectives that are trying to be achieved. Sometimes in the first article, sometimes as part of a preamble, because this shows the public that there's a clear understanding of, of the purpose of the regulation. Then when you read it, you realize that this law only targets big tech companies developing frontier models based on a revenue threshold, because I remember it was like $500 million. But then when you start. Continue reading. You realize that there are a lot of obligations and restrictions on these companies. And if you do a quick count, even with the help of AI, you're going to find out that there are 80 or 90 of these obligations or restrictions. And I'm going to say there's nothing wrong with these. They may be very bad. This law tries to prevent catastrophic events and that's a good thing to do. But we need to understand that these obligations and restrictions also impose compliance costs and burdens on these complaints. And it's a good thing, I guess, that these obligations do not kick in immediately. Some of them start until next year or 2028. But it is important for governments to assess ex antique if the cost of these requirements justify the benefits that they're trying to achieve. And you know, to be honest, we try to think of the governments as the bad guys imposing burdens on the private sector. But when you read this law, then you start seeing that it also is imposing cost on the government itself, on the government of Illinois. How? Because the Illinois Emergency Management Agency and the Office of Condor and Security and the state's attorney general will also have to comply with 10 or 12 operational mandates from this law because they have to monitor its implementation and enforce compliance. So when you think about this, the question is, are they going to receive more human, financial and technical resources to do that? And I'll give you another example. Not from the U.S. my home country, Mexico and others in Latin America in the past years have approved great laws on paper to improve regulations and tackle red tape. But always at the very end there is an article saying that the responsible agency should be doing all what the law is requiring without additional budget allocation. So this is unreasonable. You cannot have good regulations and expect them to happen spontaneously. And certainly they don't cost two pennies. So just to close this question, Brian, we need to acknowledge it's pretty sure that people and teams involved in these decisions and pieces of legislation did not find it easy. Why? Because regulating is not easy and regulating AI is even more difficult. The pace, the pressure is way different. But still, this doesn't mean that regulation cannot be a simple process. For that to happen more often, you need to have a good how to or a process to make important and everyday decisions. I guess it's going back to my home, San Fernando Mendoza. I just like high performing Correvas and athletes. Do they have a process? They tend to stick to it. Like Mendoza with the Raiders. If you have a great work ethic and a humble attitude, it's even better.
- [15:58]
B
Awesome. Thank you so much. Really appreciate that. There was a lot there to unpack. But at the end of the day you talked about the how of regulation is is really often more important and you shared many examples. But that analogy of sports teams, they have a process, a framework and of course a lot of training goes into it before they get stepped onto the field on Sunday. Good regulation, just highlighting a few things must have a good process. You need to evaluate and monitor that process. Do annual reviews of the process. I'm sure that is it still fits the mission and it's robust and you should again share a lot of examples there. But at the end of the day, regulating AI is not easy and you definitely really highlighted that part of the issue there. The next question I have for you Alfredo. What problem is your company fairly simple solving? And how can AI how can AI make regulations more accessible to everyday people?
- [16:51]
A
Thank you Brian. So I'll start saying that as a govtic startup, Fairly simple is grounded in two things. One is my personal struggle as a Mexican city who has to interact from time to time with government laws, regulations and administrative procedures. We call them tramites in Spanish. So I call them in English, the T word, because you don't want to hear that word. It can cause nightmares. And why? Because even finding regulatory information online in Mexico and many countries in the world is not simple. And once you find it in scattered government websites, understanding how it applies to you and what you have to do is even worse. So at the end of the day, you face confusion, frustration, lack of trust, and it consumes valuable time. And because of that, it's very costly. And it's not only me. This comes also from conversations I've had with firms and business associations in many countries, especially in emerging economies, where you hear similar stories. So as a lean startup, we cannot tackle all issues with regulations at once, and we cannot ask AI to do everything, to do all that heavy lifting. So what we're doing is we're starting to build a foundation based on human intelligence and expertise in law, regulation, policy, digital government, AI and machine learning. Then we're trying to focus a few core problems. One is the access of information, trying to bring access to regulations from different jurisdictions in a single point, and then use AI to help users have a clear understanding of what this implies to them. Especially they're going to be doing things for the first time, like starting a business or trying to enter a new foreign market. So let me use as an example Mexico and the U.S. so first, let's consider a scenario where you have an American firm exploring doing new business opportunities in Mexico. So I'm pretty sure they will face more regulatory struggles that I face as a citizen. And the options that they have is like, well, they could get support from DOC staff at the embassy in Mexico, and they can pay for advisors like law firms, international business facilitators, or maybe they can find a local partner in Mexico that they don't really know very well. And when you look at the other side, like considering a Mexican firm trying to enter the US Market, most likely they're not going to face the issue of accessing regulatory information because things are online and it's easy to find them. But they will face a problem of understanding a complex legal and regulatory environment that depending on their products and services, they'll need to be compliant with a lot of regulations. And if they want to operate in different states, it's like creating a different business board for each state here in the United States. And they are going to face also a country where rules are better informed and a good friend of connection is going to save the day if you get in trouble. Finally, on average, a Mexican business has less resources to pay for international advisors and fees of US Lawyers and accountants in the US So this is the space and situations where we're trying to enter as fairly simple. So we want to transform regulations into something easier to find again in a single place, easier to understand and use with the help of AI. And maybe it sounds very simple, but this is the foundation to tackle more complex problems like analyzing and simplifying larger volumes of regulations. And this is where AI brings a competitive advantage. But before, you need to have the human intelligence to understand the actual problem so you can build solutions on that and also what resonates with the market. And so if you're in a similar situation like an American affair or Mexican affair might describe, maybe in another part of the world, our vision is to reduce the regulatory friction that you face so you can make smarter and not harder decisions, especially when you're going to do things for the first time. Brian, thank you.
- [21:05]
B
Appreciate that. You again, unpacked quite a bit there around regulations. And I love how you did a comparison of the regulations in the US versus Mexico. And if people wanted to open a business, for example, in either country, some of the things that you need to go through and you need to have access to a lot of this information, of course AI allows a lot of this. But at the end of the day, what I highlighted here is you're transforming regulations in a sense, so that you can help lower that barrier to entry, whether you're starting a business or just trying to navigate that regulatory field, no matter where your business is. So I appreciate that. And Alfredo, last question of the day. If you could briefly provide, as we look ahead to the future, what do you think effective AI regulation and governance will look like over the next decade?
- [21:50]
A
Thank you, Brian. That is a great question. And my take is that if countries and organizations want to succeed in creating effective AI governance by regulation, I see that at least three things need to happen. One, this is an agenda that should not be addressed or driven only from a legal, regulatory and policy perspective. And I mean, this is not only a government agenda. So in addition to bringing politicians, lawyers, public policy folks, you also need to take people involved. You need people from the industry. And here the dynamic is not just everyone throwing elbows at each other, but cooperating in a way that is transparent, that is legitimate, and that is healthy for innovation, the market, consumers, and even the environment. I recently read that in Kansas there was earlier this year an initiative to have a task force on AI. And I understand it was a combination of the legislative branch, the executive branch at you other members outside. So I think those initiatives are great. What needs to happen sometimes is to have a real balance. So different point of views can be part of building better AI governance policy at the state level, at the country level, or even international. So a second point, every country and organizations need to take care of its regulatory house. This doesn't mean that there shouldn't be international cooperation or across levels of governance like federal, state, or even municipal. What this means is that everyone's responsibility is to have a better how to govern and regulate AI at home and not expect others to do it first. So this agenda of the how to regulate, this agenda that helped many economies achieve this current developed status. And it's the agenda that silently works in the background. Sometimes it's not seen, that's why it has been overlooked. But it's always in the background, helping other policies succeed, including AI policy. Finally, I want to mention that this agenda does not come into life and grow spontaneously. It needs to be proactively managed, coordinated, advocated by specialized institutions. In many countries, this is happening. But the last ingredient is that this agenda needs to be backed up by committed leaders. Here, the importance of leadership is that this agenda needs leaders that bring a different mindset. A mindset that in my book I called an AI regulatory humility mindset, in which you recognize that you don't have all the answers and then you need to start listening, collaborating and coordinating with others. Leaders that lead from example and not demanding things to be their way.
- [24:36]
B
Thank you, I really appreciate that. Alfredo, you talked about again, companies that want to succeed in this regulatory space. Right? It takes Paula's policymakers, lawyers, industry experts, all to come together, work on this common goal. You really broke apart policy agendas, right? They need to be proactive, robust. They need to involve all the stakeholders. It's important that these are always looking for continual improvement. Of course. And I appreciate all the examples you provided throughout our conversation today. It's just amazing with your breadth of experience in this space. And Alfredo, it was such a pleasure having you on today and I look forward to speaking with you real soon.
- [25:13]
A
Thanks for inviting me, Brian. Also looking forward to speaking again with you on your odds soon.
- [25:18]
B
Bye for now.