����vlog����

The views expressed below are those of the author and do not necessarily reflect those of the Carr Center for Human Rights Policy or Harvard Kennedy School. These perspectives have been presented to encourage debate on important public policy challenges. 

Latin America is slowly catching up to global regulatory trends on artificial intelligence (AI). While most countries in the region are still in exploratory stages, Peru has emerged as an outlier. It is currently the only country in Latin America to have enacted specific laws on AI and possesses an unparalleled number of AI-related legislative initiatives in progress. But does this legislative hyperactivity translate into meaningful regulation? A critical examination suggests not.

While legislative enthusiasm may indicate political will, it also raises concerns—particularly in contexts where institutional capacity is limited. In countries with weak regulatory institutions, AI systems risk being deployed without adequate safeguards, oversight, or redress mechanisms. This institutional fragility can amplify the harms of AI, from entrenching discrimination to enabling unchecked surveillance. As a result, the stakes of premature or superficial regulation are especially high in regions like Latin America, where inequality, corruption, and state weakness often intersect.

A Regional Outlier

Peru stands out as the most legislatively active country on AI in Latin America. Through a systematic search of official legislative databases across the region, we identified 79 AI-related bills as of January 24, 2025. Peru accounts for 17 of those, surpassing Brazil (14), Argentina (13), and Mexico (9). Importantly, Peru is also the only country in the region with two enacted AI-specific laws: Ley , promoting AI for national development, and Ley , regulating AI use in consular services.

The research focused exclusively on legislative proposals—excluding resolutions and communications—using “inteligencia artificial” as the primary search term. For bicameral systems, both chambers were reviewed. Notably, AI legislative activity in Latin America surged from 2023 onward, with 29 bills introduced that year and 44 more in 2024, reflecting a broader regional urgency to respond to the risks and opportunities of AI. Yet even in this context, Peru’s unicameral output is exceptional.

This wave of legislative enthusiasm suggests strong political interest, but also raises questions: Is Peru regulating AI with the depth and precision the technology demands—or simply legislating its promise? As we look more closely, the answer appears to lean toward the latter.

Efficiency Over Rights: The Dominant Regulatory Logic

A thematic review of Peru’s AI-related proposals shows a consistent trend: AI is predominantly framed as a tool for operational efficiency, state modernization, and technological competitiveness. From and  to public safety and judicial management, legislative texts routinely invoke the promises of optimization, productivity, and predictive precision. These algorithms, produce by technological behemoths with , are unlikely to credibly apply to Peruvians.

Bills such as  and  advertise AI’s capacity to improve transportation logistics and detect stolen vehicles. Similarly, the constitutional reform proposal  seeks to enshrine AI use as a “principle of justice administration,” with no reference to procedural fairness or algorithmic transparency. The discourse is : AI is the answer, regardless of the question.

Even in more general regulatory frameworks, such as  and  which draw heavily on the EU AI Act, human rights protections remain underdeveloped. Despite invoking concepts like “ethical AI”, “non-discrimination” and “transparency”, these principles are rarely translated into enforceable obligations. Often, they are inconsistently assigned to “users” of AI systems, rather than developers or deploying institutions, blurring lines of accountability.

The few explicit rights protections that do appear—such as the right to rectification or omission ()—are commendable though their impact remains limited in the absence of clear enforcement mechanisms. Most bills do not yet establish institutional oversight bodies, robust audit systems, or redress pathways grounded in a rights-respecting framework. While some proposals mention audits (e.g.  ) or ethical considerations (e.g. ), they are purely advisory and lack enforcement powers. These mechanisms often take a technocratic or consultative form —focused on expert opinion or voluntary principles—yet fail to embed enforceable mandates, public participation, or independent accountability.

At best, Peru is legislating the language of rights without building the institutional architecture to guarantee them.

Some of these legislative initiatives, however, are strategic in nature and may be intentionally broad, with the expectation that the executive branch will develop detailed implementation rules (“reglamentos”) after the law’s enactment. In practice, delayed publication of these rulebooks can render the law ineffective, reducing it to a mere declaration of national interest. Moreover, these reglamentos may also be subject to the political direction of each government, generating uncertainty regarding the effective implementation and continuity of legislation.  At best, Peru is legislating the language of rights without building the institutional architecture to guarantee them.

This legislative disconnect is further exacerbated by a limited technical understanding among many drafters. For example, , introduces the novel idea that “AI must not lie to a human being,” yet offers no definition of deception in algorithmic systems or how such a standard could be operationalised. Similarly,  calls for audits to detect bias in AI systems but does not specify the types of bias, responsible entities, or audit procedures. This lack of technical clarity not only weakens enforceability but also risks embedding vague or misleading assumptions into law—especially in areas where opacity and unpredictability are key features of AI systems.

Two Laws, Limited Reach

The two enacted laws offer an illustrative case.  promotes AI use for social and economic development under the banner of human-centered innovation. While the law references “ethical, sustainable, and transparent” AI and affirms respect for human rights, these references are declarative. The law’s core mission is promotional—it treats AI as a national asset to be deployed across public services, from health to defense, with no sector-specific safeguards.

is narrower in scope, applying to consular services. It mandates interoperability and digital transformation, again under the language of efficiency and ethics. This law, however, fails to define what “ethical AI” entails or to establish data protection mechanisms, procedural safeguards, or avenues for contestation. Like its predecessor, the law gestures at global standards but stops short of embedding rights-based frameworks in domestic policy.

These laws reveal a core weakness of Peru’s AI governance: it prioritizes symbolic alignment with international norms over substantive local protections. Although several initiatives reference broader frameworks on digital rights and related norms, it is not always clear whether these frameworks already contain enforceable protections or how they interact with existing legal instruments.

These laws reveal a core weakness of Peru’s AI governance: it prioritizes symbolic alignment with international norms over substantive local protections.

Who Regulates, Who Benefits?

An analysis of the actors behind this regulatory wave reveals an interesting pattern. The overwhelming majority of bills—9 out of 17—have been introduced by members of Perú Libre, a party that was constituted as a Marxist-Leninist party. Key figures like Flavio Cruz Mamani and Jhakeline Katy Ugarte Mamani are behind multiple proposals, including those on AI in security (), finance (), and justice (, ).

Yet the ideological coherence of these proposals is questionable. Rather than challenging capitalist or neoliberal paradigms, these bills often reinforce them—advancing AI as a tool for market efficiency, state control, and crime prevention, with little attention to marginalized communities or structural inequality.

Other political actors, such as Acción Popular (Carlos Zeballos, María del Carmen Alva) and Renovación Popular (José Ernesto Cueto, Esdras Medina), have also entered the regulatory race. Notably, Renovación Popular—an ultraconservative party—sponsored Law N° 31814, one of the few laws to pass.

This signals that AI is not a partisan issue in Peru; rather, it is a tool co-opted across the political spectrum for varying agendas, signaling awareness of and a desire to align with global technological trends, often without deeper ideological critique.

Of note, the bills reviewed are silent about who ought to power AI systems in the operations of the state. This is a significant oversight given the concentration of computing power and expertise within large multinational corporations, and the inherent conflict between their  and the public’s interest.

The Missing Local Lens

Many proposals show a reliance on foreign legal templates, particularly the EU AI Act and Spain’s digital transformation strategy. While comparative legal influence is not inherently problematic, it becomes an issue when local specificities are ignored. Peru, like many Latin American countries, faces distinct challenges: digital illiteracy, weak institutional capacity, regulatory capture and corruption, discriminatory policing, and fragmented public service delivery.

Some texts suggest the legislators have a limited understanding of the particular challenges of AI, referring to “” and to explainability, ignoring “black box” aspects of AI systems, such as machine learning, that render expert inspection of its programming or output anything but interpretable or transparent. And yet, AI is presented as a transparency-yielding solution to the vexing problems of corruption and lack of trust in Peruvian state action.

Yet these issues are rarely addressed in Peruvian AI bills. For instance, surveillance-focused initiatives () advocate biometric monitoring and facial recognition with no mention of Afro-Peruvian or Indigenous communities, despite well-documented patterns of . Likewise,  and  reforms fail to address unequal digital access or bias in datasets used to train decision-making algorithms.

One notable exception is , which includes a definition of “vulnerable groups” that encompasses Indigenous peoples, Afro-descendants, people living with HIV, and the digitally excluded. It also introduces the rare concept that AI “must not lie” to human users—a normative innovation worth watching. However, this bill remains stalled in committee.

Legislative Saturation vs. Regulatory Vision

Taken together, Peru’s AI governance landscape can be summarized as legislative saturation without regulatory vision. The country is prolific in drafting declarative laws but deficient in designing regulatory frameworks. The dominant approach is instrumental: AI is treated as a value-neutral tool to be inserted into state functions. Questions of who is protected, who is excluded, and who decides are sidelined.

Without robust accountability structures, these tools could entrench existing inequalities under the guise of innovation.

This is especially concerning given the accelerating uptake of AI systems in public procurement, law enforcement, and judicial administration. Without robust accountability structures, these tools could entrench existing inequalities under the guise of innovation.

Peru’s AI legislative boom offers both a warning and an opportunity. The warning lies in the ease with which AI regulation can become a performative exercise—producing laws that satisfy symbolic demands for progress, making techno-solutionist declarations while leaving structural harms unaddressed.  This exercise coupled with institutional weakness, leaves Peruvians very vulnerable to surveillance capitalism. And yet, Peru is not an outlier, but another instance of countries whose regulatory efforts cannot overcome the de facto privatization of AI, of the means for its computation, and of the goals for its use.

The way forward is clear. Peru must move from quantity to quality—strengthening the normative content of its AI legislation, institutionalizing safeguards, and opening space for public deliberation. The region is watching, and the stakes are high.

Sebastian Smart, Regional Director for the Chilean National Human Rights Institution; Technology and Human Rights Fellow, Carr-Ryan Center for Human Rights

Victor M. Montori, Robert H. and Susan M. Rewoldt Professor of Medicine, Mayo Clinic; Technology and Human Rights Fellow, Carr-Ryan Center for Human Rights