Human Rights Against Algorithmic Power
By Javier Surasky
Access Now becomes part of the AI power map as an international civil society organization that seeks to defend the rights of people and communities against technological power through advocacy, technical assistance, knowledge production, funding for grassroots organizations, and coalition-building.
The
organization was created in 2009, amid the repression that followed Iran’s
presidential elections in June of that year. At the time, it brought together
activists and technology specialists who helped restore secure communications,
protect independent websites, and disseminate images of the protests despite
government censorship.
Its work
and experience, however, expanded far beyond that moment and geography, as its
agenda also broadened. Today, its main strength lies in its ability totranslate technological harms into the language of human rights and to turn
isolated experiences of discrimination, surveillance, and exclusion into
collective demands for regulation, responsibility, and remedy. This allows us
to define it as an actor of transnational technological counterpower, one that
links the immediate protection of people exposed to digital threats with the
institutional struggle to define the terms of digital technology governance.
Deborah
Johnson (2022, p. 112), referring to algorithms, notes that “The current
discourse focuses largely on algorithms and their opacity, and I will argue
here that this reflects a narrow and inadequate understanding of
accountability,” pointing to a public debate that has paid too much attention
to the technical dimension and has not sufficiently analyzed the social and
institutional relations that emerge from these systems.
Accountability,
Johnson argues, is not only about revealing code or explaining outcomes.
Rather, “Accountability is a social practice involving actors, forums, shared
beliefs and norms, performativity, and sanctions, and the practice is aimed at
putting constraints on the exercise of power” (Johnson, 2022, p. 122).
A system
may be explainable and still remain illegitimate, for example, if it is applied
in a field where it should not be used, if it is deployed without public
participation, or if affected people do not have effective mechanisms to
challenge it. This is the space in which Access Now operates, calling for
global AI governance frameworks that are binding, enforceable, and grounded in
international human rights law.
This
explains why Access Now’s action appears both when a system makes an error and
when certain representations, populations, or forms of knowledge are pushed
aside; when the categories used reinforce stereotypes; or when people lack the
capacity to intervene in the way they are defined. This allows us to say that
algorithmic inequality is both a problem of material distribution and a problem
of representation.
Added to
this is the fact that the harms produced by automated systems are no tdistributed evenly. People living in poverty, racial minorities, migrants,
women, persons with disabilities, and other historically discriminated groups
are more exposed to automated decisions that affect them and that they may be
unable to challenge.
Virginia
Eubanks, as a case in point, analyzes how the automation of social policy can
increase the distance between those who design systems and those who suffer
their consequences. In her study on the United States, she states that “Digital
tracking and automated decision-making hide poverty from the professional
middle-class public and give the nation the ethical distance it needs to make
inhuman choices” (Eubanks, 2018, p. 13).
The author
summarizes the historical character of these infrastructures of control by
arguing that technology does not create the surveillance of poverty from
scratch, but rather updates earlier institutions and practices: “We manage the
poor as individuals in order to escape our shared responsibility for
eradicating poverty” (Eubanks, 2021, p. 25).
The same
problem appears in digital identification systems. Access Now warns that,
without adequate safeguards, they can expand surveillance, affect privacy, and
further marginalize vulnerable populations. Its position is not to reject all
digital identification, but to demand that an assessment be made first of whether the system is necessary and whether less intrusive alternatives exist.
That is why
we say that Access Now functions as a counterweight to a form of governance
that pushes human and social concerns into the background, pressing for
affected people and communities to occupy the center of technological
governance.
Niamh
Kinchin refers to the insufficient consideration of rights as a “procedural
gap”: “Without procedural justice and ‘voice’, defined as consent,
transparency, and a right to be heard, algorithmic decisions will fail to find
legitimacy and acceptance” (Kinchin, 2024, p. 1).
By
including the reference to “voice,” a critical political dimension of AI
governance becomes visible: an impact assessment does not, by itself, guarantee
that affected communities can influence the design of a technology or prevent
its deployment. Here, a key function of Access Now appears: contributing to the
creation of spaces where the experiences of affected people can be transformed
into regulatory arguments.
Access Now
helps expose a structural asymmetry in the AI power map. Companies have access
to internal information, technical teams, and resources that allow them to
influence legislative processes on an ongoing basis, while social
organizations, with considerably fewer resources, must simultaneously cover
multiple jurisdictions and respond to agendas largely defined by governments
and corporations.
This
element is aggravated by the absence of even the slightest trace of democratic
order in the ecosystem of AI actors: “Users did not elect Zuckerberg to govern,
nor do they have any representation in Big Tech’s private governance”
(Srivastava, 2023, p. 991).
Access Now
also intervenes in the face of this legitimacy gap by demanding that powerful
actors exercise the basic obligation of due diligence, conduct impact
assessments, ensure transparency, provide grievance mechanisms, and accept
prohibitions on uses that are incompatible with the enjoyment of human rights.
What has
been said allows us to organize Access Now’s participation through five
components:
1. The
normative component, which it produces when it turns technological problems
into legal categories.
2. The
coalition component, which emerges from its work with people and local
organizations and in international networks that carry and amplify the demands
of the former, creating cross-cutting spaces of concern in response to claims
that would otherwise remain fragmented.
3. The
technical component, which has allowed Access Now to develop a Digital Security
Helpline to respond to attacks, surveillance, malware, and other threats
against journalists, activists, and human rights defenders, while also
generating knowledge about the strategies and functioning of digital
repression.
4. Documentation,
which adds to the above through Access Now’s participation in campaigns such as
#KeepItOn to end internet shutdowns by recording internet disruptions and their
consequences. This, in turn, helps produce evidence, identify those responsible,
and promote international advocacy actions (Rosson & Anthonio, 2026).
5. Funding,
through the provision of flexible grants to grassroots organizations working to
defend human rights in digital environments, redistributing part of its
resources toward actors with local knowledge.
Taken
together, Access Now’s proposal goes far beyond the creation of voluntary
frameworks or ethical declarations, prioritizing binding rules and enforcement
mechanisms aligned with human rights and respectful of local populations and
communities, especially vulnerable populations.
Its
strategy appears to be twofold: to defend prohibitions when a technological use
is incompatible with rights and, at the same time, to strengthen the safeguards
applicable to systems that remain permitted. This combination of positions
leads Access Now to act within institutions, but also to push for changes in
the boundaries of what those same institutions consider acceptable, giving
greater weight to the consideration of people’s rights or, in other words,
disputing the political and legal conditions under which AI may be used.
At a
historical moment in which “Algorithmic accountability is in the making” (2022,
p. 123), Access Now seeks to identify responsible actors, create avenues for
challenging algorithmic decisions, strengthen local communities in the face of
an AI that is global and unequal by nature, and promote the idea that abuses
taking place in the field of AI should not go unpunished.
Basic data
- Access Now was created by Brett Solomon, Cameran Ashraf, Sina Rabbani, and Kim Pham in 2009 to help restore secure communications, protect independent websites, and disseminate images of the protests that followed that year’s elections in Iran and faced harsh government censorship. Today it is an international civil society organization dedicated to defending and expanding the digital rights of people and communities, especially those vulnerable to AI-related harms.
- It is headquartered in California, but has a stable presence in New York, San José, Costa Rica, Brussels, Berlin, Tunis, Delhi, Nairobi, and Manila.
- It promotes binding, enforceable governance frameworks based on international human rights law through strategies that combine political advocacy, technical assistance, funding for grassroots organizations, research production, and the articulation of international coalitions.
- Its Digital Security Helpline provides free, direct, and ongoing technical assistance to journalists, activists, media outlets, social organizations, and human rights defenders.
- RightsCon is one of its main international convening spaces on human rights and technology.
- In 2015, it was nominated for the Human Rights Tulip Award by David Kaye, then UN Special Rapporteur on freedom of opinion and expression
References
Eubanks, V.
(2018). Automating inequality: How high-tech tools profile, police, and punish
the poor. St. Martin’s Press.
Eubanks, V. (2021). La automatización de la desigualdad:
Herramientas de tecnología avanzada para supervisar y castigar a los pobres (G.
Deza, Trans.). Capitán Swing.
(Original work published 2018).
Johnson, D.
G. (2022). Algorithmic accountability in the making. Social Philosophy and
Policy, 38(2), 111–127. https://doi.org/10.1017/S0265052522000073
Kinchin, N.
(2024). “Voiceless”: The procedural gap in algorithmic justice. International
Journal of Law and Information Technology, 32, eaae024. https://doi.org/10.1093/ijlit/eaae024
Rosson, Z.,
& Anthonio, F. (2026). Rising repression meets global resistance: Internet
shutdowns in 2025. KeepItOn. https://www.accessnow.org/wp-content/uploads/2026/03/KeepItOn-Internet-Shutdowns-2025-Annual-Report.pdf
Srivastava,
S. (2023). Algorithmic governance and the international politics of Big Tech. Perspectives
on Politics, 21(3), 989–1000. https://doi.org/10.1017/S1537592721003145