—Dr Elif Biber, Legal Scholar in European Public Law and Digitalization, University of Luxembourg
A New Question?
How do fundamental and human rights emerge, and what conditions shape their emergence in the digital age? In this post, I aim to reflect on a puzzling social phenomenon: the widespread sense of comfort – or at least resignation – that individuals display toward phone applications, AI systems, and algorithmic decision-making tools that process personal data and shape life opportunities.
This comfort stands in sharp contrast to how we react to interference in the physical world. Being watched by a person or a camera in a public space typically provokes discomfort, or at the very least a sense of awareness. Yet comparable forms of monitoring, profiling, and behavioural influence through digital systems often pass without scrutiny. Users rarely question how applications process data, whether they manipulate behaviour, or how persuasive and personalised digital environments constrain autonomy. This disparity raises foundational questions: Why do we tolerate digital interference so readily? What does this tolerance reveal about evolving expectations and vulnerabilities surrounding our rights? And do existing human rights frameworks adequately capture the nature of power in the digital realm?
I argue that they do not. In particular, I contend that the emergence of human rights in the digital environment is structurally constrained by what I call the phantom influence of digital technologies. Drawing on an Arendtian understanding of rights, I suggest that the architecture of advanced AI systems restricts the socio-political conditions under which rights traditionally emerge. This calls for renewed conceptual and institutional responses.
Before turning to Hannah Arendt, it is necessary to situate the discussion within broader debates on the nature of human rights. As John Tasioulas asks, “What are we talking about when we talk about human rights?” His 2012 account identifies three significant approaches: the reductive view, which grounds human rights in terms of a vocabulary of human interests; the orthodox view, which treats human rights as moral rights grounded in humanity itself; and the political view, which defines human rights through their function within international legal and political practice.
This post does not seek to critique these theories directly. Rather, its aim is to ask whether something distinctive about the digital realm challenges their assumptions. Specifically, my aim is to explore a central question: can the ‘rule of the algorithm’ open up space for new digital human rights to emerge, or does this rule instead erode the socio-political foundations from which rights have traditionally arisen?
An Arendtian Account of Rights: Bringing the Phantom Influence of Artificial Intelligence Systems to Light
To address this question, I turn to the legal and political philosophy of Hannah Arendt, whose critique of abstract human rights remains highly relevant in the contemporary age. Writing in response to the adoption of the Universal Declaration of Human Rights, Arendt famously argued that “the world found nothing sacred in the abstract nakedness of being human.” For her, rights do not arise from mere humanity, nor are they guaranteed by formal legal recognition alone. Instead, rights emerge in the space between law and lawlessness, through the capacity to act in a shared public realm. As Arendt maintained, freedom is not an inner state but a political practice: to act is to initiate something new, to speak and appear before others within a web of human relationships. Rights, on this view, are not static entitlements but political achievements grounded in action, plurality, and contestation. This understanding has profound implications for digital environments. If rights depend on visibility, contestability, and collective action, then the conditions under which digital power operates become legally and politically significant. I argue that the architecture of advanced digital technologies systematically restricts the capacity to act in the Arendtian sense. This restriction does not arise primarily from overt coercion but from structural features of digital systems that obscure power relations and suppress political contestation. I identify three such features: virtuality, complexity, and dynamism.
Virtuality
Digital platforms and AI systems are inherently virtual, non-physical, and de-territorialised. In contrast to the physical realm, where state actors exert significant influence, the virtual environment is not primarily governed by territorial control. As a result, social and political activities within these environments lack the tangible presence that enables resistance, visibility, and accountability in the physical world. Put simply, monitoring by a person or a camera in a physical environment is not comparable to monitoring conducted through a phone application. Within online platforms and AI systems, the experience is fully mediated by digital interfaces, reshaping how surveillance is perceived and enacted. This virtuality significantly weakens the experiential conditions necessary for recognising the concrete interference with fundamental and human rights. Power becomes ambient, mediated, and difficult to localise.
Complexity
Digital systems are also characterised by profound complexity. Many AI applications function as “black boxes,” producing outcomes without intelligible explanations. Even relatively simple algorithmic systems, such as those used in welfare administration, can be extremely difficult to understand or contest for users, in particular vulnerable individuals. It is also important to emphasize that, despite superficial similarities in output, humans and AI systems rely on fundamentally different processes to achieve results. As Joseph Weizenbaum warned decades ago, collapsing the distinction between human and machine reasoning risks reducing human agency to mechanistic processes. Furthermore, the complexity ofsuch systems is amplified by deliberate design choices that increase opacity, including manipulative interface designs known as dark patterns. Furthermore, recent controversies, such as Meta’s decision to use Europeans’ personal data for AI training without prior consent —despite GDPR requirements— underscore how these characteristics can undermine users’ agency. According to a NOYB-commissioned Gallup survey of 1,000 Meta users in Germany, while nearly 75% were aware of Meta’s plans, only 7% actually supported the use of their data for AI training, leaving an estimated 68 million people across Europe entirely unaware of the change.
Furthermore, the complexity of advanced technological systems is compounded by the phenomenon of “AI hype”, which frequently inflates public perceptions of AI while obscuring its real limitations. This dynamic is critically examined by Arvind Narayanan and Sayash Kapoor in their book AI Snake Oil: What Artificial Intelligence Can Do, What It Can’t, and How to Tell the Difference, where they use the term “AI snake oil” to describe a marketing-driven practice in which companies overstate AI’s capabilities to sell products or services that ultimately provide limited value. Such exaggeration not only fuels aggressive and misleading marketing narratives but also encourages an unhealthy overreliance on AI systems, a trajectory that may pose serious long-term risks to humanity. In this context, AI-style hallucinations in judicial environments, generative AI within education and its implications for the democratic project, and the emergence of a post-literate culture in which thinking becomes a luxury highlight the urgent need to prioritize critical reasoning and human judgment over technological dependence.
Dynamism
Finally, digital technologies evolve at extraordinary speed. According to the Stanford AI Index, the computational capacity for AI has been doubling every few months, far exceeding the pace predicted by Moore’s Law. This rapid evolution destabilises legal categories and renders regulatory responses perpetually reactive. For instance, the emergence of large language models (LLMs) in late 2022 exposed the limits of the EU AI-Act’s risk-based approach, as these general purpose systems were not envisaged in the 2021 AI Act Proposal.
The Phantom Influence
Taken together, these features render the tension between individual liberty and private power far less visible than in the physical world. Even pervasive abuses may go unrecognised. I describe this phenomenon as AI’s phantom influence: a form of power that operates through virtuality, complexity, and constant transformation. Phantom influence arises from the architecture of digital systems and diminishes the impact of core socio-political factors that would otherwise render the tension between individual freedoms and private power visible or tangible. It operates in the space between private corporations that design digital technologies and the individuals who use them and it concerns the suppression of conditions necessary for the articulation and exercise of rights. The tension between freedom and lawlessness is not eliminated but obscured, “phantomized”.
However, it should be underlined that digital technologies can also empower rights. Social platforms have created alternative public spaces that enable expression, association, and political mobilisation, particularly in authoritarian contexts, where such platforms may provide the few remaining venues for open political discussion. In these settings, digital spaces can, in practice, provide more accessible and effective arenas for political dialogue than physical places such as cafés, bars, or even parliamentary chambers. As Habermas has pointed out, however, these platforms are not neutral extensions of the public sphere. They are commercial infrastructures whose logic is tied to the commercial exploitation of networked communication.
Habermas’ observation helps to identify two significant dimensions when discussing human rights in the context of new advanced technologies: (i) Using technology as a tool to exercise rights differs fundamentally from (ii) asserting rights against technological power itself – such as demanding transparency, contesting automated decisions, or resisting surveillance. In the latter case, technology is not merely an instrument but the party of rights claims. It is here that phantom influence is most acute.
Therefore, the political dimension represented by the phantom influence remains markedly weak. Although private actors are the primary creators of digital technologies, they cannot be regarded as fully fledged political actors operating within a genuinely pluralistic framework. It is important to reiterate that this analysis does not concern the first case outlined above, in which technology functions as a tool and private creators may assume an explicit political role. Rather, the focus here is on the second case. In this latter scenario, the “language” of these systems – that is, their code – is largely inaccessible and unintelligible to individuals. This structural opacity profoundly undermines the socio-political dimension of rights and gives rise to unprecedented conditions for the exercise and abuse of real power. This situation also explains why many digital rights claims – often framed as an “Internet Bill of Rights” – emerge outside traditional political processes, frequently driven by civil society actors rather than mass political mobilisation.
These dynamics expose also the limits of the normative equivalency paradigm, which holds that rights enjoyed offline must simply be protected online. While normatively appealing, this approach risks obscuring the distinctive forms of power exercised by digital systems. Virtuality, complexity, and dynamism may require not only new interpretations of existing rights but also the recognition of new ones, such as a right to a human decision. Without accurately identifying how digital interferences undermine our autonomy, legal protection may remain blind and ineffective.
Strengthening the Conditions for Rights
If the phantom influence weakens the socio-political conditions for rights, how might its effects be mitigated? I briefly highlight two promising avenues. First, civil society plays a crucial role. NGOs have been central in documenting AI-related abuses and advocating for public interests, as recognised in the 2022 report of the Center for AI and Digital Policy. Initiatives such as “Reclaim Your Face,” which calls for a ban on indiscriminate biometric surveillance, demonstrate how collective action can reintroduce visibility and contestation into digital governance. Second, judicial interpretation remains essential. Landmark cases such as Google Spain, decided by the Court of Justice of the European Union, illustrate how courts can adapt existing rights to new technological realities. As Aharon Barak argues, rights must be interpreted in light of both their historical purpose and contemporary conditions if they are to remain “living” norms. Contemporary conditions of fundamental and human rights may also require considering inter-legal approaches, as they are recognized across numerous legal systems.
Conclusion: Toward a Primavera Digitale
Reflecting on the phantom influence of digital technologies reveals the necessity for intentional and creative action—action that not only mitigates its negative effects but also actively cultivates a meaningful digital environment in which genuine, critical engagement can flourish. Strengthening civil society participation and cultivating rights-sensitive judicial interpretation can help restore the conditions under which rights may meaningfully emerge. I envision this space as a Primavera Digitale: a digital spring inspired by Botticelli’s vision of renewal. In such a space, individuals are not passive data subjects but active political agents. The internet, reclaimed as a civilisation of the mind rather than an instrument of dehumanisation, can once again become a domain of speech, creativity, and action – where the capacity to act is not phantomized but empowered.
Suggested citation: Elif Biber, A Challenge for a Primavera Digitale: The Phantom Influence of Artificial Intelligence Systems, Int’l J. Const. L. Blog, Feb. 3, 2026, at: http://www.iconnectblog.com/a-challenge-for-a-primavera-digitale-the-phantom-influence-of-artificial-intelligence-systems/