Privacy Talk with Thiago Guimarães Moraes, PhD candidate in a joint-degree program between Universidade de Brasilia (UnB) and Vrije Universiteit Brussels (VUB): How does Brazil work for responsible AI?
“This interview has been recorded on 16th July 2024 and discusses AI and data protection legislation.”
- How does Brazil work for responsible AI?
- What is the challenges with ethical guideline?
- How does Brazil work for responsible AI?
Thiago: Yeah, okay, yeah. First of all, this goal of achieving responsible AI, it’s very relevant and at the same time, very challenging. And we can see that governments from all over the world are putting a lot of effort into doing that.
We can say for sure that Japan, for example, is leading a lot of initiatives. So as we’ve seen when they promoted the G7 the data free flow with trust, and now, more recently with the Hiroshima protocol and strategy regarding the generative AI, so what we see is that this discussion has not started today.
It has been happening for a while, and thankfully, Brazil is also very concerned on this issue. But of course, there are several interests that come into force, and we have to be very careful and assessing what are the legitimacy of these interests and how they could. How can we raise some kind of consensus.
And this always takes a while. And why saying that? Because we have different initiatives in Brazil, I briefly mention the ones that I consider more relevant, and maybe you understand my point. I believe the most.
The first leading initiative from Brazil regarding this topic was its Brazilian National Strategy. You know the AI strategy. So the Brazilian National AI strategy, which we call EBIA.
（Video: When Brazil discusses AI and work, diversity has a seat at the table）
It was launched in 2020, so actually, by the time I was back in Brazil, and I worked in the drafting of this strategy, unfortunately, for several political reasons, it couldn’t develop. It couldn’t mature as we wanted it to.
It had for political reasons, it had to be published as it was. I am afraid it was published a bit incomplete, in the sense that there were still some goals or milestones that needed to be better estimated, but in any sense it was important so we could position ourselves right.
And why I mentioned this strategy? Because of this strategy, of course, it was looking for economic productivity. So it had several axes looking for the educational sector, the productive sector and work related.
And at the same time, there was some transversal access that was looking for ethics, governance, regulation. So this was important because it was a document that was trying to find this balance already, and I think this document has influenced the discussion that then followed up in the legislative.
Because in the Brazilian Congress, the first legislation that calls more attention on AI regulation. It was very principiological in the sense that they were going to have some principles, and then we would live for self regulation.
You know, we’ll leave for companies to do as they seem fit. They could. They should look for this principles and just go and this actually connects with many initiatives that we’re seeing other other places at the time, like the EU, which were like ethical or ethical guidelines and stuff like that, which, of course, is a very important first step.
But we know that in the end, it’s a very weak way of, like, guaranteeing that things are going to happen, because you just have to live to trust that people will follow these ethical guidelines.
- What is the challenges with ethical guideline?
And be even more challenging if these ethical guidelines are not translated how they’re going to be implemented, the things can become very chaotic, because what’s transparent for me may not be what’s transparent for you, what’s fair for me may not be the same idea of fairness for you, and especially for other minorities.
So of course, we have to find some mechanisms that make this ethical approach more achievable, implementable, and enforced, and that’s when the discussion in the Congress, the Brazilian Congress, starts to develop.
So when the discussion reached the Brazilian Senate, they created a commission legal expert, which did have a very throughput benchmark in the OECD countries, to look how things were developed in these different regions.
Then they start to draft a new bill. It’s the bill that’s being currently discussed. So this commission of legal experts drafted the first version. Then another Commission was created with senators.
And these commissions, all of those, held several public hearings with representatives from all sectors, from different economic sectors, from corporate authors, from civil society, academics, and little by little, we are developing this legislation that is trying to guarantee, of course, rules that can force their innovation.
So it has rules for initiatives such as sandbox. It has rules for giving some flexibility for startups because, of course, they can’t have a high regulatory barrier, so they have to have some flexibility.
It also has initiatives to promote that the executive branch has to further develop the educational system to include more computation and data science in the curriculum of students.
But at the same time, it’s also legislation that’s looking for some governance mechanisms, for some rights that needs to be guaranteed, and for governing all that we have been trying to find this balance of not having just one central authority, because definitely that would be very challenging for this authority to work with such a transversal technology.
So we want to have a sector approach, which in some way remembers what remind us what we’re seeing in the US, but looking also at the European experience, we’re looking for some kind of coordination.
So like the European will have the AI board with all the authorities, what we’re doing at the national level.
We are proposing that all sectoral regulators like the telecom regulator, transportation regulator, the energy regulator, all of those may use AI, and all of those will enforce these rules in their sectors, but we’ll have a coordination system with a central coordinator authority just to guarantee that this collaboration and some harmonization happens.
So this is the proposal as of now. Of course, there are still some ways to go. And in Brazil, sometimes things can go, just go 180 degrees back. You know, we know there’s a lot of resistance from some sectors, especially some tech companies. So we cannot say that everything is assured for now, but this is the way that we are paving and we are developing our debate.
Kohei: That’s a very exciting time to put the directions to where we should go in terms of the AI period, which is the board within this new technical innovation over the war. The problem is how we can take a balance in between innovations, protection, so the worries are the key part of the actions.
All the regulators, who are the stakeholders and involvement in the processes and the as for the next question might be fits of these topics, because you mentioned about the sandbox. I mean, sometimes that’s a part of the good initiative, such as in Singapore, which is working in a very proactively put in the regulators and sandbox system is a privacy tech space.
So my question is, what is the future role of the regulatory sandbox in Brazil? So if you have anything, just like any practical use cases, some kind of thing that could be helpful. Could you share your ideas?
Thank you for reading and please contact me if you want to join interview together.
Privacy Talk is the global community with diversified expert, and contact me below Linkedin if we can work together!
