09/02/2021

In an interview with Sputnik News Agency, Imbuzeiro discusses the use of AI in Brazil.

Imagem: Freepik

With potential applications in diverse fields of knowledge, artificial intelligence has become a topic of public debate in Brazil, which is discussing the creation of a legal framework for the development and use of AI by the public and private sectors in the country. In an interview with Agência Sputnik , IMPA researcher Roberto Imbuzeiro spoke about Brazil's potential in using this technology and the ethical and legal implications that must be developed for the discussion around the topic to mature.

"Artificial intelligence is based on computing methods and the availability of data. So, all sectors where there is data are sectors where you can use AI," he explained. According to the mathematician, areas in which Brazil leads, such as agribusiness, paper production, and healthcare, are natural sectors to introduce the technology.

According to Imbuzeiro, the legal framework for AI, which has not yet been scheduled for a vote, should address ethical and social issues related to the use of data and the decisions made by AI. “There has been talk of racism in artificial intelligence methods. There was a story where Google image search recognized photos of Black people as if they were monkeys. This isn't exactly a problem with the method, it's a problem with the data. Probably, Google engineers didn't train the photo recognition system with enough data on Black people. And, like this one, there are several other related cases in the world, involving racism, sexism, involving other forms of injustice, perhaps.”

Read the full article :

Brazil could 'leap ahead' in areas of AI application, according to experts.

Although it lags behind other countries in the technological race, Brazil has the potential to excel in some areas of artificial intelligence application, believes researcher Roberto Imbuzeiro, from the Institute of Pure and Applied Mathematics.

In 2019, Brazil adhered to the recommendations of the Council of the Organisation for Economic Co-operation and Development (OECD) on artificial intelligence (AI). On February 4, 2020, Bill No. 21/2020, which creates the legal framework for the development and use of artificial intelligence by the public and private sectors in the country, was presented to the Chamber of Deputies. And, a year later, the bill is still awaiting the appointment of a rapporteur in the Committee on Science and Technology, Communication and Informatics (CCTCI).

Artificial intelligence systems can be applied to a wide variety of fields of knowledge, in any sector that works with data. This is explained by mathematician Roberto Imbuzeiro, a researcher at the Institute of Pure and Applied Mathematics (IMPA).

"Artificial intelligence is based on computing methods and the availability of data. So, any sector where there is data is a sector where you can use AI. In principle, the more data, the better," he states in an interview with Sputnik Brazil.

According to the expert, these systems basically try to find patterns in the data and take action based on them. A good example, he cites, is Facebook. This social network takes all the information relating to a particular user, establishes a relationship between "their data and their likes". From there, the service presents ads and "things to pique their curiosity or their desire to consume" and "that's how Facebook makes money".

"We see that Brazil has, for example, leading companies in agribusiness, paper production, healthcare, and imaging exams. In short, we have a wide range of sectors in which Brazil is very well positioned, including on the international stage. And these sectors are, let's say, natural sectors for you to increasingly introduce artificial intelligence."

Despite the numerous benefits attributed to this technology, what has been widely discussed lately, according to the mathematician, are ethical and social issues in general related to the use of data and the decisions made by AI.

"There has been talk of racism in artificial intelligence methods. There was a story where Google image search recognized photos of Black people as if they were monkeys. This isn't exactly a problem with the method, it's a problem with the data. Probably, Google engineers didn't train the photo recognition system with enough data on Black people. And, like this one, there are several other related cases in the world, involving racism, sexism, and perhaps other forms of injustice. And this, in fact, needs to be addressed by legislation. But I think that, perhaps, the topic is not yet mature. Nowhere in the world is there a consensus on exactly what to do, who is responsible, and what happens."

These are the types of problems that should be addressed in the AI legal framework, which is not yet scheduled for a vote.

"It's a debate that's happening right now, it's something we need to understand better, and perhaps it's not yet the right time to codify these things into law, which is always a serious step. Of course, it has the advantage of protecting people. But it also makes things quite rigid."

Despite Brazil's great potential, according to Imbuzeiro, Brazil is currently far behind other countries in this technological race, which is dominated by the United States, China, and Canada.

"But we can do things, within our scale and using our data, that can be very useful. Not only internally, for our market, for our people, but also externally. They can become pioneering technologies. So, we are behind, but I see a lot of opportunity for us to leap ahead, at least in some areas of AI application."