By Javier Surasky
.png)
The Prior Architecture of the Digital Experience
In the
international geopolitics of AI, we have traditionally looked for power in
rules, enforcement capacity, infrastructure, and computing power. But in the
digital environment, there is a way in which power is expressed less visibly
and operates as a structure that precedes decision-making: the initial
configuration of systems.
When we
open a device, use a browser, or run a search, we do not start from scratch. We
enter a prior architecture in which someone has decided what options appear, in
what order, in what language, and with what level of visibility and associated
costs for not using them.
The order,
description, colors, size of the buttons, and the moment when an option appears
influence user behavior.
From there,
we can think of a “geopolitics of digital defaults” that shapes the terrain
from which choices are made, guides our experiences, reduces friction for some
paths and increases it for others, with the result that following certain
options appears more likely than following others.
The default
has a special force because it rests on very ordinary behaviors: few people
review the settings of every device or program they use, or read every privacy
policy they accept, just to move forward.
In their
book El algoritmo paternalista, published in 2024, Ujué Agudo Díaz and
Karlos Liberal remind us that “we avoid decisions whenever possible, especially
when they are difficult and complex. For that reason, the most attractive
alternative is usually to accept the predetermined choice, which is known as
the default effect” (pp. 135–136). In other words, we are most likely to take
“what is already installed,” or what is presented as the path that requires the
least expenditure of time, applying a predictable rationality that designers of
digital systems and devices take advantage of by giving us systems prepared to
work quickly, but also incomplete information.
The initial
configuration is much more than a detail meant to help the user. It is the way the
company welcomes the user and presents the option it prefers them to adopt. It
can make the difference between achieving a dominant market position or
remaining practically invisible, between a normalized data-collection practice
and one that requires active consent.
The power
of the default lies in the user not perceiving it as a decision. Instead, the
user naturally takes it as the system’s contribution to their convenience, even
though behind every predetermined option lie hierarchies, economic interests,
cultural assumptions, and power relations.
Defaults, Markets, and Regulation
In the
European Union, the 2022 Digital Markets Act (DMA) does not expressly mention
default-option design, but it does restrict design and interface practices that
may take advantage of user inertia through preselected options or exit
barriers. It requires that certain functions, such as unsubscribing from core
platform services, be available under clear and non-obstructive conditions,
thereby reducing the possibility that user choice will be conditioned by
artificial friction rather than by a free decision.
The
regulation echoes the power of certain platforms that act as gatekeepers to
critical digital services, such as search engines, app stores, browsers, and
operating systems, and seeks to open spaces of choice for users and companies
that depend on those ecosystems.
For the
first time, this regulation addresses the design of defaults as a problem that
begins in design but extends to competition, consumer rights, data protection,
and market power. When a platform controls the operating system, the app store,
the browser, the search engine, the payment system, the assistant, and more,
the default option serves as a position of structural advantage by capturing
the starting point.
Apple, for
example, was forced to introduce changes for European Union users after the DMA
entered into force, including a choice screen for the browser to be used, a
separate section for default applications, and options that had not previously
existed to modify or delete them.
We see
something similar in the case against Google over the search market in the
United States, where the Department of Justice argued that distribution
agreements placing Google as the default search engine on devices and browsers
helped maintain its dominant position, bringing visibility to distribution
contracts, agreements with manufacturers, and the default placement of
programs.
Behind
these decisions lies an economic element that drives initial designs, because
being the “default option” means receiving traffic, data, learning, scale, and
familiarity. For that reason, every user who remains with the initial
configuration reinforces the dominant service; every query, interaction, and
preference feeds data into systems that improve through use.
But it
would be a mistake to think that this is merely a problem of economic
advantage, because a geopolitical perspective, like the one proposed here,
shows how default options distribute authority and create dependency.
A default
search engine can condition the information that becomes prioritized, or even
accessible; the use of a map involves cartographic and territorial
position-taking; the default language favors certain markets and populations,
leaving others behind; a privacy setting alters companies’ ability to access
data; and the list goes on.
In that
context, the question of defaults comes close to a broader question of digital sovereignty.
Generative AI as Cognitive Mediation
Generative
artificial intelligence only intensifies the scale of the problem of default
options.
With AI
assistants, the default begins to operate as a layer of cognitive mediation
that gives the user one answer among many possible ones or, in other words,
orders the information the user will receive without the user being aware of it
or having been given an explanation.
Technical
questions, such as which model answers, with what sources, within what cultural
framework, under what criteria of relevance, and with what limits, acquire
great political weight.
In
addition, with AI the interface stops showing lists of possibilities and
instead produces a single formulation. This is easily visible in search
engines, where the integration of artificial intelligence has taken us from a
list of links (trapped in a bubble, but at least still a list) to a faster,
more practical, and more direct answer, but a single one. The default shifts
from the field of services to the field of access to information, producing
uniformity of responses and shielding the bubble that already concerned us
because it was, in itself, an obstruction to dialogue and the diversity of
views.
There is
also a legal dimension that we must consider, because the initial configuration
can “push” the user toward a particular jurisdiction, certain terms of service,
a specific dispute-resolution system, or particular privacy standards. In
transnational digital environments, accepting a configuration means accepting
the applicability of a legal corpus associated with it through contracts and
internal policies that users do not read, nor are prepared to understand: the
full Windows agreement, in English, is more than 17,000 words long, slightly
longer than the English version of The Little Prince; Spotify’s
agreement, in Spanish and for Argentina, has more than 6,100 words, while
Borges’s short story “Funes el Memorioso” does not reach 3,000.
Conclusions
Default
designs largely simplify matters of usability, security, compatibility, or
efficiency, since every device, application, or assistant needs some point of
reference as a starting point.
The geopolitical question is who defines that point, where it is defined, and with
what effective possibility of modification. The lack of transparency about what
comes preconfigured, the reasons behind it, and the consequences of accepting
it are evident to any user, worsening a situation that is already serious if we
accept that “the default” has social effects.
The
geopolitical dispute over AI is, among many other things, a competition to
establish the default values of the digital ecosystem. This means leading the
definition of the initial decision-making environment, placing the struggle to
establish default options at the center: whoever prevails will make their
proposal appear as the “natural” alternative for other governments, companies,
and users, and their technical-decisional architecture will end up conditioning
norms, markets, and strategic alignments.
For that reason, it is necessary to deepen mechanisms for auditing default options that are capable of identifying whether default configurations are affecting rights, competition rules, access to public information, or citizen participation, and to make their operating mechanisms transparent, under the premise that every predetermined interface is an opportunity to organize social relations, distribute opportunities, and show or hide.
Digital geopolitics is a dispute over the sovereignty of knowledge, information, and access to data: in an increasingly digital society, whoever designs digital normality exercises power, and whoever defines the default establishes the point from which others begin to decide. I do not think that in the twenty-first century we should call that soft power, but rather “coloniality by default.”