False Start: Peru's AI Regulation Celebrated Without Being Read

View on LinkedIn

On September 9, 2025, Peru published Supreme Decree No. 115-2025-PCM, implementing its Artificial Intelligence Law. The immediate reactions were predictably enthusiastic: "Peru leads AI regulation in Latin America!" "A pioneering regulatory framework!" "A model of technology governance!"

But there is a problem: those celebrating do not appear to have read it.

The "first country to…" syndrome

Latin America has a particular fascination with being "the first" to regulate emerging technologies. The desire for leadership is understandable, but regulating technology we neither understand nor use is building castles in the air.

This decree reveals its disconnect from the very first definitions:

AI: "A scientific discipline oriented toward the research and development of AI-based systems"

This is a tautology. Defining Artificial Intelligence using "AI-based systems" is like defining medicine as "the science that studies medical things." It explains nothing. A student who wrote this would fail. Did no one at the Office of the Presidency of the Council of Ministers notice they were defining something by using that very thing?

AI-based system: "A machine-based system that, for explicit or implicit objectives, infers from the input it receives how to generate outputs such as predictions, content, recommendations, or decisions that may influence one or more physical or digital environments."

This definition reveals a limited understanding of AI, focused almost exclusively on generative systems like ChatGPT. What about computer vision systems? Classification models? Optimization networks? The definition reads as if written by someone who discovered chatbots and assumed that was all there was to AI.

Regulating ghosts: the Peruvian AI regulation nobody needed

But here comes the revelation that makes everything even more bewildering: this regulation applies primarily to Public Administration entities. The private sector is excluded, except when it participates in the National Digital Transformation System. We are therefore faced with a fascinating paradox: the Peruvian State has created 36 articles of exhaustive regulation to oversee its own artificial intelligence systems. Which systems? That is the million-dollar question that apparently no one asked during the months of drafting. The Peruvian government is not developing language models, has no computer vision projects, and is not building advanced predictive systems. The reality is more prosaic: the Peruvian State can barely keep its websites running, and now must comply with algorithmic impact assessments for systems it simply does not possess. The implicit confession on every page

Every article of the regulation is an involuntary confession of the true state of Peruvian government technology. When they devote paragraphs to regulating facial recognition they do not have, they are admitting they aspire to have it someday. When they classify risks of algorithmic credit-scoring systems, they reveal they have not even digitized the basic processes that would precede any algorithm. When they create a sandbox for experimentation, they confess there is nothing to experiment with. It is an aspirational document disguised as a regulatory framework, a plan for what they would like to regulate if they ever had anything to regulate.

The celebration that reveals everything

And yet, LinkedIn feeds fill with celebrations. "Peru leads AI regulation," proclaim those who read no further than the headline. Consultants are already preparing compliance courses that nobody needs. Academics are organizing conferences on a regulatory framework for nonexistent technology. It is a complete ecosystem built around a void, an economy of simulation where everyone pretends that regulating is as good as building, that having a regulation equals having the technology. The celebration itself is the most compelling proof of our problem: we prefer the appearance of progress to actual progress, the theater of modernity to modernity itself.

Supreme Decree No. 115-2025-PCM is not a regulatory framework; it is a mirror. And what it reflects is not a country ready to lead in AI, but one that confuses decrees with development, regulation with realization, and paperwork with progress. A country that needs fewer regulations for technologies it does not have, and more honesty about the capabilities it needs to build.

Recommended complementary reading: https://www.hks.harvard.edu/centers/carr-ryan/our-work/carr-ryan-commentary/perus-ai-regulatory-boom-quantity-without-depth