Platforms as Sites of Experimentation

Two verdicts in the United States in March 2026 demonstrate that courts are increasingly willing to treat online platforms as designed systems rather than neutral conduits or editorial voices. In Los Angeles, a jury found Meta and Google liable on product-design and failure-to-warn theories for building platforms that are addictive and thus dangerous to minors – evaluating not what content circulated but how the system was engineered. In New Mexico, a jury found Meta liable for misleading users about the safety of its apps, violating state consumer protection laws. The legal theories in these cases differ, but their underlying logic is the same: the tech companies’ legal shields do not protect their technological designs.

Yet this turn to platform design in litigation requires identifiable plaintiffs and causal chains, and does not directly address the deeper question of who should govern this design, and with what transparency and accountability mechanisms. Article 15 of the International Covenant on Economic, Social and Cultural Rights (ICESCR), particularly the human right to science, provides a normative framework for addressing this gap. Specifically: Article 15(1)(b) protects the right to enjoy the benefits of scientific progress and its applications; Article 15(2) addresses the conservation, development, and diffusion of science; and Article 15(3) protects the freedom indispensable for scientific research. Together, these provisions offer a normative framework for determining who may participate in shaping the scientific and technological systems that structure public life.

For the three decades spanning the passage of Section 230 of the Communications Decency Act in the US to newer regulations such as the EU’s Digital Services Act (DSA) and the UK’s Online Safety Act, the policy and regulatory debates on how to regulate online platforms and what responsibilities they bear have largely focused on civil and political rights, mainly freedom of expression. Even where regulatory instruments extend beyond speech and cover other rights such as privacy, dignity, non-discrimination, and freedom of information, they remain centered on civil and political rights. Less attention is given to economic, social, and cultural rights (see observations here, here, and here). Most notably, the human right to science has rarely been seriously put forward when examining online platforms. This omission is consequential. More than just technological products that apply scientific knowledge, online platforms are socio-technical systems that continuously experiment on human behavior and social interaction. Through large-scale data extraction, behavioral experimentation, and engagement optimization techniques, platforms learn from, predict, and influence human attention and behavior. Beyond this, platforms increasingly mediate access to scientific knowledge itself. Yet the platforms’ technological designs are left largely in the hands of a small number of private corporations. This makes the human right to science a forgotten right in platform governance, depriving scholarly and regulatory debates of the obligations that this right generates.

The Forgotten Right

The human right to science, enshrined in Article 15 of the ICESCR, has received little attention and has long remained on the margins of international human rights law. Some rights related to science, such as freedom of thought and expression, academic freedom, and freedom of association, have been considered justiciable civil and political rights. Conversely, the “right to science” itself has been bundled with the economic, social, and cultural rights – also sometimes called “second generation rights” – which some scholars deem less justiciable for being more dependent on available resources and progressive realization through State measures. While the Vienna Declaration and Programme of Action rejects the Cold War-era division between civil and political rights, on one hand, and economic, social, and cultural rights, on the other, this artificial divide has impacted the development and realization of the right to science.

The neglect of this right in platform governance debates has profound consequences. The discussions are almost always centered on matters involving expression: content moderation, censorship, editorial discretion, and the marketplace of ideas. Even when scholarly and regulatory debates expand beyond freedom of expression and include other rights, the analysis remains within the civil and political rights paradigm. The technological systems themselves – ranking algorithms, engagement mechanics, behavioral optimization – are almost never examined as scientific advancements in which the public holds participatory rights under international human rights law. Accordingly, neither public participation in the design of these systems nor independent research access to their operations is recognized as an entitlement arising from the human right to science.

Tools of Scientific Experimentation 

More than conduits of users’ speech, platforms are instruments of scientific experimentation: they use applied mathematics, data science, and behavioral psychology, and deploy algorithmic systems and engagement mechanics to learn from, predict, and influence human attention and behavior at scale. Because platforms engage in ongoing large-scale experimentation, everyone has the right not only to enjoy their benefits (through utilizing them to exercise freedom of expression) but also to “participate” in them. This means that decisions regarding platforms’ technological systems – including those that determine whose attention is worth monetizing, whose knowledge circulates, and whose participation in public life is amplified or suppressed – should not be left to a small group of elite corporations.

As Samantha Besson argues, science is a participatory good that gives rise to a set of participatory and collective rights to contribute to and enjoy the benefits of science. That the right to science has a participatory dimension and is not limited to the passive enjoyment of science’s benefits is supported by the UN Committee on Economic, Social and Cultural Rights’ General Comment No. 25 (2020), interpreting the right to science enshrined in ICESCR. It is also supported by Article 27 of the Universal Declaration of Human Rights, whose official French text speaks of “le droit… de participer au progrès scientifique” (rendered in the English version as “the right… to share in scientific advancement”). To operationalize this right in the context of online platforms, the platforms’ technological systems must not only be transparent – which some regulations already require – but also explainable, subject to public scrutiny, and open to meaningful public participation in their designs.

General Comment No. 25 has already moved in this direction by bringing these systems within the remit of Article 15. In interpreting the right to science and emphasizing that decisions on emerging technologies must be taken within a human rights framework, General Comment No. 25 called on State parties to the ICESCR to “develop mechanisms so that autonomous intelligent systems are designed in ways that avoid discrimination, enable their decisions to be explained, and allow accountability for their use.” It also underscored that State parties must “establish a legal framework that imposes on non-State actors a duty of human rights due diligence, especially in the case of big technology companies.” What General Comment No. 25 has not yet fully developed is how to operationalize the participatory dimension of the right to science with respect to these companies’ technological systems.

A Human-Right-to-Science Framework in Platform Governance

Applying a human-right-to-science framework in platform governance requires democratic participation, sharing the benefits of scientific progress, and freedom of scientific research. Consequently, if the right to science is taken seriously in platform governance, States must adopt at least three measures: participatory governance mechanisms for affected stakeholders, structured researcher data access, and algorithmic auditability.

Democratic Participation. More than a right to passively receive technology, the right to science includes a right to actively participate in deciding how scientific advancements should contribute to society. When a platform unilaterally designs ranking algorithms, amplification logic, and engagement incentives to learn from and influence its billion users, it monopolizes the decisions over a participatory good – one in which the public holds a right not merely to benefit, but to participate. Thus, unlike the right to democratic participation under Article 25 of the International Covenant on Civil and Political Rights, the right to science generates the right to participate in the scientific architecture of platforms’ systems, not just their political outputs.

Platform appeals mechanisms, such as the Oversight Board, the DSA’s out-of-court dispute settlement bodies, and India’s Grievance Appellate Committee, while providing remedies to users, are inadequate: they allow users to contest individual content decisions but leave the underlying architecture entirely beyond the public’s reach. Genuine Article 15 compliance requires meaningful participation in the design of this architecture, not merely the adjudication of what flows through it. Systemic-risk assessments (see DSA’s Article 34), while a significant step towards evidence-based oversight of platforms, are likewise insufficient: aside from not guaranteeing transparency on platform design’s relation to risks, they do not provide a legal basis for continuous access to and participation in platform design.

States should therefore require platforms to establish participatory governance processes that bring scientific communities, civil society, and affected publics into decisions about platforms’ technological designs. This requires public participation in managing the risks arising from the technological architecture itself. Given that platforms’ algorithms reinforce social drivers and can have neurophysiological impact on users, decisions regarding their design – including how they optimize attention, predict and shape behavior, and serve certain epistemic interests – should consider stakeholder input and be subject to public oversight. Such participation need not entail direct co-design by users, but could include, for instance, independent oversight structures, institutionalized consultation, and representation of affected stakeholders in governance processes. The DSA already gestures towards this model, but grounds it on systemic-risk regulation, rather than in the human right to science. This difference matters: a right-to-science grounding transforms these features from contingent regulatory mandates into fundamental human rights requirements. One concrete institutional model would be a platform design council, akin in structure to social media councils, but focused on overseeing platform design.

Sharing the Benefits of Scientific Progress. Dominant platforms extract enormous scientific value – knowledge about human psychology, social dynamics, and attention economics – from the collective behavior and data of their users. A right-to-science framework demands that this knowledge not be locked into exclusive private control but be subject to public scrutiny via meaningful transparency and accountability mechanisms.

To ensure this, States should mandate structured and legally protected researcher data access. Article 40 of the DSA points toward this model, but grounds researcher access in the need for systemic-risk research. A right-to-science grounding would transform it from a mere policy choice to a human rights requirement, enforceable before domestic courts as a right States must protect. This also means that the human right to science provides a normative basis for researchers outside the EU to claim access to relevant platform data within their jurisdictions.

Extending this argument further, States can also mandate researcher data access as a precondition for the liability immunities that platforms currently enjoy. This would prevent platforms from hiding behind these liability immunities to shield their technological designs from independent scrutiny. Technological infrastructure that extracts, learns from, and even experiments on user data must be subject to public audit.

Research Freedom. Article 15(3) protects “the freedom indispensable for scientific research.” This generates the researcher data access argument in its strongest form. When platforms restrict API access, impose prohibitive terms of service on academics, or make independent audit severely compromised, they impede the scientific study of systems that profoundly affect public life. States bear a positive obligation under Article 15(3) to protect this freedom – not as a mere policy choice, but as a human rights requirement.

Translating this obligation into the machinery of digital regulation requires algorithmic auditability as a condition of operating epistemic infrastructure at scale. A system whose operations cannot be meaningfully studied by independent scientists is incompatible with this obligation.

Remembering the Forgotten

The human right to science under Article 15 of the ICESCR generates positive State obligations that existing regulatory and analytical approaches to platform governance, centered on civil and political rights, have overlooked. The right-to-science framework shifts our focus from reacting to harms after they occur to demanding public participation in the design of our digital architecture – grounded not in contingent policy choices, but in human rights.

The right to science has been forgotten in the digital sphere for too long. Recovering it widens our legal imagination in addressing the structural challenges posed by online platforms that continuously experiment with and shape human behavior.