Home » Latin American lawmakers pushing for AI regulation – but to what extent?
America Business Economy Featured Global News Latin America National Security News Politics South America Technology

Latin American lawmakers pushing for AI regulation – but to what extent?

Amidst the fast growth of generative AI and natural language tools, Latin American countries are moving to create norms and to update current frameworks to respond to the use of these new technologies.

In line with movements such as that of the European parliament, which has just approved a bill that regulates the use of AI in the EU, lawmakers in various Latin American countries are beginning to either update existing legislation on less advanced technology or create new frameworks for AI.

In Nicaragua, representatives of the national assembly participated on Thursday in a “virtual parliamentary exchange on artificial intelligence and digital rights” held by the Inter-American parliamentary network ParlAmericas, and agreed to create a taskforce to develop proposals for AI regulation.

ParlAmericas comprises representatives from 35 legislatures in North, Central and South America.

According to Nicaraguan lawmaker Jenny Martínez, a digital parliamentary group will work on a common agenda on artificial intelligence, digital rights of individuals and user protection. Nicaragua itself already has a law focusing on cybercrime that Martínez believes could be a good template.

In Mexico, lawmaker Ignacio Loyola, from the country’s conservative opposition party PAN, presented a bill to “regulate the use of artificial intelligence and robotics for governmental, economic, commercial, administrative, communicational and financial [purposes] so that its use is always based on adherence to ethics and adherence to law.”

The bill was presented two weeks ago in congress and is yet to start its passage through the legislature.

In Brazil, senate president Rodrigo Pacheco submitted last a month a bill, a result of recommendations made by lawyers and jurists in a working group created in 2022, centered on five central topics for AI: principles, the rights of those affected, risk classification, governance obligations and requirements, and oversight and accountability.

The proposal brings together various bills previously presented in congress into a single regulatory framework for AI. The initiative also foresees launching regulatory sandboxes.

Meanwhile, Colombian ICT minister Mauricio Lizcano announced in recent days the creation of an artificial intelligence laboratory in the country with the goal of prototyping technologies and regulations for those technologies.

“The great challenge we have is protection, privacy and data analysis, the black box to identify and control the algorithm. Another challenge is to ensure that AI is diverse, equitable and can be within everyone’s reach,” Lizcano was quoted as saying.

From a global coordination perspective, UN Secretary-General António Guterres said during the week that he was in favor of the idea of creating an international agency for artificial intelligence as part of the UN system. 

Guterres said this institution could be molded on the International Atomic Energy Association, which promotes the peaceful use of nuclear energy, and have scientific and research purposes.

Caracas-based development bank CAF, for its part, said in an article last month that the concern about how the new generative AI tools will affect Latin America and the Caribbean “is reasonable, especially if we consider the concerns that already exist in the region on levels of employability and digital literacy.” 

The bank said the unemployment rate in Latin America is already high and more than 25% of jobs in the region may be at risk of being replaced by automation, according to a 2020 OECD report.

While backing AI regulation, CAF said actions that respond “disproportionately to the hype and fear that currently surrounds technology should be avoided.”

The bank said banning the technology would limit the region’s ability to influence the design of new AI tools and to “modulate” their deployment on a regional scale. 

CAF recommends that countries in the region prioritize less specific regulations on AI and adopt regulatory sandboxes to help measure the impact of proposals.

There is also the issue of having some sort of common ground across legislation – a topic that is not new when it comes to discussing AI, as reported by BNamericas in 2019.

There is a general consensus among those working with AI on the need to create guidelines for the technology.

Most of the tech companies operating in different stages of the chain with whom BNamericas has spoken in recent weeks are in favor of regulation, although to varying degrees. 

“I think it will be inevitable [AI regulation] – and it will come. It must come. In the past there has been a lot of debate about the obstacles data protection laws could bring. It turned out to be an excellent approach for people to have data security,” Cristina Palmaka, SAP president for Latin America, told BNamericas.

“But at SAP, we moved beyond the buzzword to enter the application of this generative AI in the world in which we are specialists, which is the corporate world,” she said during SAP’s Sapphire event in São Paulo earlier this month.

According to Palmaka, SAP has an ethics committee for AI development. The executive lists as key topics intellectual property, data security, social and racial biases, and being clear on what is real and human-made and what is not.

Among other initiatives, SAP announced at the event that it will use ChatGPT, from Microsoft-backed OpenAI, in its HR management platform SuccessFactors.

The two companies are expected to collaborate on integrating SuccessFactors with Microsoft 365 Copilot, the company’s AI resource, and Copilot in Viva Learning, a hub centered on Microsoft Teams.

Supply chain, HR and customer relationships, in addition to back-office platforms, are the first use cases of SAP’s systems powered by generative AI. 

“But it won’t stop there”, said Palmaka.

Source: bnamericas