The data agreements that feed algorithms
President Luis Abinader signed Decree 787-26 on 2 of November 2026. The decree includes as an annex the Code of Ethics for Artificial Intelligence in Public Services.

President Luis Abinader signed Decree 787-26 on 2 November 2026. Attached to it is the Code of Ethics for Artificial Intelligence in Public Services. The text binds centralized and decentralized institutions that design, procure or use artificial intelligence, and its coverage extends down to a public servant's writing assistant.
In April 2025 we wrote that «driven by the Artificial Intelligence fad, the authorities started backwards». The 787-26 repeats the same pattern.
For regional context, the Latin American Artificial Intelligence Index 2026 from ECLAC and Cenia presents Chile as the regional leader with 71,4 points out of 100. The Dominican Republic ranked 10 out of 20 countries, with 45,1. In October 2025, when the country dropped from seventh to ninth place, we described the result as «absolute progress accompanied by relative decline». In governance we scored 73,2, and in research, development and adoption barely 27,1. This sample sharpens the reality that we write rules more easily than we can enforce them.
In its structure, Decree 787-26 prohibits 7 uses, among them social scoring and criminal profiling by neighborhood or economic condition, with the DIGEIG being the institution that must approve the evaluations of high-impact systems.
No fines of its own, none. If we cross the Atlantic Ocean, we find that Europe fines violators up to 7% of global turnover. It seems to us a good code, but with little force to be enforced.
What we have been saying about data
The most unsettling thing about the decree is that it leaves what has already been done without consequence. We know that article 110 of the Constitution orders that the law only provides and applies for the future.
With that in mind, in August 2024 we had warned that «the mistake is looking for the fever in the bedsheet». The fever is now artificial intelligence. At that time we called for strengthening «the framework of respect for digital citizenship, starting with the government itself».
In 2013, the JCE opened its master registry of ID holders to public entities, banks and companies, with written authorization from the holder; two years later it agreed with Migration on capturing fingerprints and photos of everyone who enters or leaves the country.
In 2024, 33 institutions established a connection framework through the X-Road platform in a security pact. «Interoperability», some proclaimed. The structure linked the DGII, the TSS, Public Health and the three electric distributors, which enabled automatic alerts when a wanted person requests a service. As of this writing, no public portal gathers those agreements.
Since July 2026 data exchange within government has been a general obligation, under Decree 403-26, which ratified the machinery for processing citizen data.
Since December 2013, the Ley 172-13 treats the transfer of personal data without the data subject's consent as unlawful. The agreements were built under that law, so reviewing them would be applying law that was already in force, with no retroactivity at all. Let us recall that the law exempts only investigation and intelligence agencies, so we doubt an electric distributor fits that exception.
High-impact systems already in operation must undergo an evaluation within the period set by the DIGEIG. The process should establish which data enter and through which agreement they arrived. All institutions have 180 days to comply, a period that would close around 2 May 2027. A prohibited profiling practice that remains active after that will be a new violation.
What we ask of the DIGEIG
We demand «human rights impact assessments before implementing public digital systems», as we argued in April 2025; also transparency about who handles government data. That is why we ask the DIGEIG for a public registry of the data agreements in force, with the parties, the date, the data that circulate, the purpose, the legal basis and the automated systems that consume them. The Ley 200-04 on Free Access to Public Information supports the request since 2004.
The National Council on Ethics of Artificial Intelligence is responsible for evaluating the Code each year. We hope its first report can tell us how many institutions published their registry. We hope it says how many active flows enabled on X-Road have a published agreement with an identified legal basis. The danger is not in artificial intelligence, but in the ethics of bureaucrats in their use of citizen data.
Originally published in Acento.