–Luis E. López Rodriguez, Venezuelan lawyer and LL.M. graduate of Loyola University Chicago School of Law
In November 2025, weeks before his arrest and extradition to the United States, Nicolás Maduro urged Venezuelans to use VenApp, a state “citizen attention” application repurposed for political surveillance, to report “everything they see and hear.” By January, Maduro was in U.S. custody and Venezuela was governing itself under a state of exception. The app did not leave with him. Neither did the carnet de la patria, the ZTE-assisted identity card that links political data to food and benefit access for millions of people, nor the Sistema Patria payment platform through which much of the population receives subsidies and pensions. Whatever government consolidates in Caracas, it will inherit not only captured courts and a contested amnesty, but a running apparatus: databases that remember who marched and who informed, scoring systems that decide who eats first, and an application through which citizens were taught to watch each other.
The constitutional question is not whether this kind of digital surveillance is compatible with democracy. That question has a mature literature, what Steven Feldstein and others call the global expansion of AI surveillance, and an easy answer. The constitutional problem is what happens when the apparatus survives the autocrat. This post argues that authoritarian legality now exists in two forms, written and executable, and that transitional constitutionalism possesses tools only for the first. Every technique a new democracy can deploy against inherited authoritarian law, repeal, narrow construction, value-reinterpretation, desuetude, presupposes a norm that acts through human interpreters. Code has no interpreters. It has operators. And the gravest danger is not that a successor regime will abuse the inherited apparatus in bad faith, but that a well-intentioned one will keep running it in good faith, because it works.
Two forms of authoritarian legality
By executable legality I mean two overlapping things. First, rules embedded in operational state systems, identity gating, eligibility scoring, automated benefit allocation, citizen-reporting platforms, that apply themselves continuously, without needing to be invoked, argued, or construed. Second, the data layer beneath those systems: the political profiles, informant reports, and behavioral records that an authoritarian regime accumulates and that constitute, in machine-readable form, its institutional memory of who is loyal and who is not.
This is not the familiar debate about whether states may automate governance. It is a question that debate has not asked: what is the constitutional status of automation inherited from a regime whose defining feature was the fusion of allocation and political control? Transitional constitutionalism has a rich toolkit for inherited rules and inherited personnel. It has nothing for inherited systems.
Why transitional doctrine cannot reach code
Consider what transitions know how to do with authoritarian law, and watch each technique fail.
A statute is inert until applied; repeal it, and it stops. A system is not inert: it runs until someone decommissions it, and repealing its authorizing law changes nothing about its operation, assuming an authorizing law exists at all. As I have argued in related work on algorithmic delegation, these systems typically enter the state tacitly, through decree, procurement contract, or nothing at all, leaving no legal form for a successor to repeal. The Venezuelan social-control architecture was never enacted; it was deployed.
A court can construe an inherited statute strictly, refuse to extend it by analogy, read it against the new constitutional order’s values, the move the Czech Constitutional Court made when it accepted the continuity of norms but declared a discontinuity of values with the old regime. There is no algorithmic equivalent of strict construction. A scoring model has parameters, not text; it cannot be read narrowly, only run or not run. The interpretive techniques on which all post-authoritarian jurisprudence relies presume a gap between norm and application in which judgment can operate. Executable legality’s defining feature is the elimination of that gap.
And law’s harms can be stopped prospectively; data’s cannot. A repealed statute ceases to bind, but a database does not cease to know. The regime’s archive of political behavior is a form of capture that survives any institutional reform, because it can be copied, queried, and weaponized years later by whoever controls it, including, crucially, by a future government of any stripe.
The comparative record
There is one great precedent, and it proves both that the problem is solvable and that solving it requires institutional invention. Germany did not handle the Stasi files through ordinary repeal or judicial interpretation; it created a dedicated statute and an independent custodial authority — the Stasi Records Archive, that preserved the archive for accountability and victims’ access while walling it off from operational use. But the Stasi archive was paper: inert, finite, dead. Today’s inheritances are operational and self-updating. Ecuador’s Chinese-built ECU-911 surveillance network, with its documented feed to the domestic intelligence service, passed intact from the government that installed it to its successors. Tunisia inherited Ben Ali’s internet-censorship machinery in 2011, and proposals to reactivate it resurfaced within years. The pattern across cases is consistent: successor governments keep the apparatus, because it is useful, because it is how pensions get paid and streets get watched, and because no constitutional doctrine tells them they may not.
Venezuela now poses the question in its hardest form. The removal of a leader, I have argued, does not decompose the regime he built; and the succession underway since January 3 is unfolding inside the institutional shell of the regime it succeeds. The digital layer is the sharpest instance. A transitional government that uses carnet data to distribute food efficiently, or VenApp to manage an emergency, is not exiting Chavismo; it is administering Chavismo’s nervous system in democracy’s name. This is capture by inheritance, and it will be done sincerely, by officials who believe they are stabilizing a country, which is precisely what makes it durable.
Digital lustration
What would a constitutional response look like? Transitions learned long ago that they must vet persons: lustration. I want to suggest that transitions must now learn to vet systems. Call it digital lustration. Three principles give it content.
First, sunset by default. State digital systems inherited from an authoritarian period should presumptively lapse unless expressly reenacted by statute, after public deliberation, within a defined transitional window. Nothing governs by inertia. Reenactment forces the new order to take deliberate ownership of whatever it keeps — and it forces disclosure, because a legislature cannot reenact what no one can describe.
Second, custodial separation. The regime’s accumulated political data should pass to an independent custodial authority on the German model: preserved, because victims have a right to know and prosecutors will need evidence; but subject to a strict non-use rule, data collected by the old regime for political control is inadmissible in any state decision allocating rights, benefits, or burdens.
Third, non-migration. Inherited datasets and models may not seed successor systems. A democratic state may build digital infrastructure, it almost certainly must, but it builds on data its own legal order generated, under its own constitutional constraints.
The objections are real. Transitional states are typically poor, and the inherited platform is often the only functioning rail for payments and identity: but sunset-with-reenactment permits keeping the plumbing after deliberate democratic adoption; what it forbids is silent continuity, and what custodial separation removes is the profiling layer, not the pipes. One might also object that this is technology policy, not constitutional law: but when state systems allocate rights according to political data, what is at stake is equality, due process, and the boundaries of delegable public power, the constitutional core, not its periphery. And the German precedent might suggest ordinary legislation can solve the problem after the fact: but the German statute governed a dead archive, while an operational system compounds its harms every day it runs, which is why the default must be constitutional and must bind before ordinary politics discovers how convenient the inheritance is.
We have learned to ask what new democracies should do with the autocrat’s judges, his statutes, even his constitution. We have not learned to ask what they should do with his servers. The most faithful servant of a deposed autocrat no longer waits in the barracks or sits on the bench; it runs quietly in a data center, allocating, scoring, and remembering. A transition that rewrites the constitution but keeps the code has changed the text of the law while leaving its execution in the old regime’s hands.
Suggested citation: Luis E. López Rodriguez, When the Algorithm Survives the Autocrat: Executable Legality and the Digital Inheritance of Democratic Transitions, Int’l J. Const. L. Blog, Jul. 7, 2026, at: http://www.iconnectblog.com/when-the-algorithm-survives-the-autocrat-executable-legality-and-the-digital-inheritance-of-democratic-transitions/