Former Chilean President Gabriel Boric Font speaking at the official launch of Latam-GPT in Santiago, Chile. Image: Alex Ibáñez / Prensa Presidencia de la República de Chile
Chile is emerging as a regional leader on AI, with the country spearheading the launch of Latam-GPT in February. It’s the first open-source AI language model trained on data from Latin America. Then-President Gabriel Boric said it was ‘strategic and urgent’ that Latin America plays a part in the development of AI.
A key question for many countries is whether to take regulatory inspiration from the EU or the US – or neither.
A bill on AI currently before Chile’s Congress is one of several recent legislative initiatives either proposed or enacted in Latin America as the region grapples with the challenges and opportunities presented by the technology. Peru implemented a law in 2023 – updated in 2025 – that classifies AI systems by the level of risk posed to human rights and civil liberties. For example, the use of AI in workplace recruitment or employee dismissal is classified as high risk, requiring human oversight, record-keeping and algorithmic transparency. Similar draft legislation has been put forward in Brazil, while El Salvador enacted an AI law in 2025.
Jose Luis Ambrosy
Officer, IBA Latin American Regional Forum
The approach to AI governance seen so far in Latin America follows the EU model, says Jose Luis Ambrosy, an officer of the IBA Latin American Regional Forum. Patricia Silberman, who’s also an officer of the Forum, also highlights that the region’s ‘personal data protection and cybersecurity regulations – matters related to AI – are more similar to the EU system than the US one.’
Ambrosy describes the current regional agenda across Latin America as being to adopt a ‘horizontal, risk-based framework inspired by Europe, with a focus on fundamental rights.’ This model classifies AI systems according to their level of risk – for example, unacceptable or limited – which then creates specific obligations for developers. The US model, in contrast, favours governance that ‘reduces rigidity, allows for rapid iteration and focuses resources on where there is the greatest real risk,’ believes Ambrosy, a partner with Claro y Cia in Santiago.
Others have highlighted the benefits of the EU approach to AI regulation. For example, in a paper, the NGO Carnegie Europe says that the EU’s AI Act reflects the bloc’s ‘long-standing approach of prioritising high ethical standards and fundamental rights in tech and digital policies – a strategy of fostering both excellence and trust in human-centric AI models.’
Critics of the US approach at the federal level, meanwhile, have noted that by focusing on lighter-touch regulation, obligations in areas such as disclosure and liability may be less than concrete.
There’s been some debate, therefore, as to whether countries in Latin America should adopt AI governance structures modelled after the EU example or consider alternatives. There are concerns about how easily Latin American countries would be able to implement ‘EU-style’ legislation, given they may lack the resources possessed by the bloc’s Member States.
Andrea Saffie, an officer of the IBA Latin American Regional Forum, says the risks stemming from following the EU model include the possibility of ‘over‑regulation’ that hinders innovation. The ‘extensive obligations’ found under the EU framework – relating to risk governance, technical documentation, testing and incident reporting, for example – ‘can strain regulators with limited resources’ as well as SMEs, she says, meaning innovation is pushed ‘underground’ into the informal sector.
Chile’s bill was first put forward in spring 2024. At that time, Chile’s Transparency Council, an independent public law body, warned in connection to the bill that ‘excessive or premature regulation’ could harm innovation and development. It also highlighted an issue with ‘insufficient transparency/explainability – particularly when AI is used in higher‑risk contexts such as the public sector.’
Ambrosy says that Chile’s bill now includes, for example, measures to support innovation – including regulatory ‘sandboxes’ allowing AI innovations to be more safely tested – as well as steps to protect confidentiality. As such, ‘it addresses the innovation-protection dilemma, incorporating sandboxes and measures for SMEs alongside reinforced obligations for high-risk uses,’ he says.
Take-up of AI in Latin America has so far been moderate compared with the US, parts of Western Europe and Asia. The technology is mostly used in the finance, commerce/logistics, customer service, marketing, health and agriculture sectors.
Organisations such as the Inter-American Development Bank have identified a number of specific challenges for Latin America regarding its adoption of AI, such as gaps in data privacy regulation and the existence of unregulated markets for buying customer data, as well as the risk of surveillance, for example.
Saffie, a partner at Cariola Diez Perez-Cotapos in Santiago, says a further key challenge is the ‘regional legal heterogeneity’ of Latin America. ‘The EU starts from a harmonised baseline,’ she says, referencing the bloc’s Single Market, while in Latin America, ‘differences in personal data, consumer, cybersecurity and competition law require fine‑tuned adjustments to avoid gaps or overlaps.’
She also highlights that Latin America and Europe have ‘different priority sectors’ when it comes to AI. In Latin America, the greatest value from AI will come from ‘applications that address everyday frictions in areas such as financial inclusion and payments,’ she says, while in the EU, AI-related R&D and deployment is aimed at ‘highly regulated, precision industries’ such as advanced manufacturing, medical devices and energy. So implementation should also be different.
In Chile, another legislative proposal before the country’s Congress would ‘regulate the creation and dissemination of realistic digital imitations of people’s images, bodies or voices generated by AI,’ says Silberman, a partner at Carey in Santiago. She adds that the proposal is particularly concerned with copyright issues and deepfakes, which are closely linked to AI. But, Silberman suggests the new Chilean government’s priorities ‘do not appear to include this more precautionary project’ among the immediate areas of focus.
International Bar Association 2026 © Privacy policy Terms & conditions Cookie Settings Harassment policy
International Bar Association is incorporated as a Not-for-Profit Corporation under the laws of the State of New York in the United States of America and is registered with the Department of State of the State of New York with registration number 071114000655 – and the liability of its members is limited. Its registered address in New York is c/o Capitol Services Inc, 1218 Central Avenue, Suite 100, Albany, New York 12205.
The London office of International Bar Association is registered in England and Wales as a branch with registration number FC028342.
