Opening Technology / Democratizing Society: The Ethical Imperative in the Age of AI

For the 2025–2026 academic year, I had the privilege of serving as a Research Fellow at the Center for Hellenic Studies, Harvard University. It was a fascinating and transformative experience, and I consider it one of the most prestigious and significant fellowships currently available to researchers in Greece.

As a CHS Research Fellow in AI Ethics, I pursued my central research question through multiple complementary perspectives, including the philosophy of technology, ontology, epistemology (both modern and contemporary), the political philosophy of technology, and the relationship between artificial intelligence and democracy. During the fellowship period, I published several essays presenting these different dimensions of my research, while additional works are currently forthcoming. The publications produced during my CHS fellowship are the following:

In what follows, I outline the central research question that shaped my fellowship and examine how I addressed it through the body of work I produced during my time at the Center for Hellenic Studies.

In my research, I focus on the concept of openness. I explore whether AI – as a field, not just as a set of applications – can become more open. This question led me beyond the technical understanding of “open source” toward a broader investigation of openness across different domains of life: ontological, socio-political, and technical. My work suggests that meaningful openness in AI requires engagement across all three domains. In the ontological domain, I draw on the work of Henri Bergson; in the socio-political domain, on the field of the commons; and in the technical domain, on thinkers such as Andrew Feenberg, as well as constructivist approaches in Science and Technology Studies (STS) and open-source practices. Bringing these perspectives together, I argue that socio-technical spheres should not be treated as separate, but as processes that co-emerge and co-shape one another. From this perspective, ethics is not simply a matter of regulation or a set of guidelines for responsible use. Rather, it becomes a way of opening up new possibilities – for both society and technology – starting from one of the most influential emerging fields today: Artificial Intelligence. In this sense, my work approaches ethics through the lens of philosophy of technology, particularly within the tradition of critical theory.

The central research question that guided my CHS fellowship was: What would ethical and open AI look like from a commons-based design perspective? In the process of addressing this question, I found it necessary to ask a more fundamental one: What kind of openness do the commons suggest? This, in turn, raised broader questions about the kinds of societies and technologies that such forms of openness co-produce. These inquiries led me toward an ontological investigation of the nature of openness – a concept that has since become a central thread running through all of my subsequent work.

We can divide the notion of openness found in literature in some categories, therefore developing a kind of “typology of openness”: First, corporate openness (and in particular AI-corporations), as described by Nick Bostrom, which demonstrates that even apparently transparent exchanges of information between companies exist within institutional and legal constraints that may limit genuine collaboration. Second, Open Science initiatives, intended to democratize knowledge and promote open access publications, confront structural challenges. Although the ideals of Open Science are compelling, in practice they often require substantial financial and institutional resources, producing uneven accessibility. Third, FLOSS communities provide practical models of openness through collaborative and technically oriented practices, though participation depends on both social and technical literacy. Fourth, STS scholarship demonstrates the historical, epistemic, and social construction of technological artifacts, making visible the mechanisms and assumptions that might otherwise remain opaque. Fifth, Feenberg’s theory of the democratization of technology underscores technology’s empowering potential, but highlights that such potential is contingent upon enabling social and institutional conditions. Each of these perspectives contributes to understanding openness, yet also reveals limitations when considered in isolation. A sixth and comprehensive understanding of openness emerges most clearly in the context of the (digital) commons – an original (both theoretical and practical) political approach to social, technical, organizational, epistemological and economic, thus in one word, ontological conditions – where openness is not merely the sharing of resources but the structuring of interwoven social, ethical, and technical infrastructures that sustain collective knowledge creation and technological design, participation, and ongoing institutional reflection.

So, to begin assembling the puzzle, societal openness complements technological openness. Societies that cultivate participatory norms, transparency, and shared technological infrastructures are more resilient, equitable, and ethically coherent. Legal or regulatory measures alone are insufficient; without attention to ontological, social, and institutional dimensions, such interventions risk reproducing closure, inequality, and ethical blind spots. Ethical AI therefore requires alignment between technological, social, and political systems: design, governance, and institutional practices must collectively support reflection, experimentation, and participatory engagement. Only through such integration can the democratic and ethical potential of AI and other technologies be fully realized.

The first puzzle piece is the political philosophy of Karl Popper and Cornelius Castoriadis, which reinforces the open political dimension: imaginary institutions, norms, and social imaginaries must support open, continuous engagement, preventing closure of knowledge, authority, or ethical reflection. Popper’s concept of the “open society” situates political freedom as inseparable from epistemic openness, emphasizing institutional designs that allow criticism, accountability, and adaptive change. Castoriadis, on the other hand, foregrounds the creative and self-instituting nature of social imaginaries: societies and thus technologies are not pre-given, but continually created and recreated by collective action, imagination, and normative innovation. Together, they demonstrate that technological systems and AI cannot be ethically evaluated in isolation from the political and normative frameworks that shape their deployment and social effects; and that they need to be evaluated continuously (therefore openly) rather than being enclosed to only a few “evaluators” – corporations.

On top of these, the second puzzle piece the philosophy of technology – particularly the theory developed by Andrew Feenberg – provides crucial conceptual tools for understanding and guiding ethical AI. The study of AI serves as a prism through which we examine foundational ontological assumptions, social norms, and political structures. Ethics, in this framework, mainly by Feenberg, is not a narrow concern with compliance or design choices but a broader project of cultivating conditions that allow societies and individuals to act, reflect, and participate in shaping their technological and social worlds. Opening technology is therefore inseparable from democratizing society: it is a method and a goal, guiding design, governance, and social engagement toward continuous processes of learning, reflection, and ethical deliberation.

As a final comment here, AI, due to its planetary scale and rapid deployment, magnifies these concerns. Its technologies, infrastructures, and social consequences operate across local, societal, and global scales simultaneously. Ethical evaluation must therefore address immediate effects, longer-term systemic impacts, and planetary implications. Open AI systems, designed with these scales in mind, should become instruments of reflection, participation, and democratization rather than mechanisms of control. By considering AI in this broader context, we recognize that ethical AI is inseparable from social, political, and ontological considerations. It is only within this integrated framework that design, governance, and participation can co-evolve to sustain ethical, democratic, and reflective practices.

The third puzzle piece is process philosophy, as informed by Bergson and A. N. Whitehead (or even some thoughts from Heraclitus, the ancient Greek philosopher) which provides a metaphysical grounding for understanding openness. Process philosophy emphasizes becoming, relationality, and continuous evolution. Openness, in this sense, is not merely procedural but ontological: it is the condition under which creation, reflection, and participation can flourish. Technologies like AI operate within these dynamic flows, capable either of reinforcing closure or facilitating openness. Political ontology, drawing from Popper and Castoriadis, highlights that technological systems are embedded in social imaginaries, norms, and institutional frameworks. Closures are enacted not just through technical mechanisms but also through social, ethical, and political practices that consolidate power, restrict knowledge, or inhibit participation. Openness requires alignment between ontic actions (design and implementation) and ontological principles (normative and relational assumptions), enabling continuous adaptation, experimentation, and democratic engagement.

Through this kind of analysis we encounter what I call the “double ontological openness”. This means the creation not only of technical or societal openings (ontic), but also the ontological conditions for the continuation of the search for openings. This dual approach ensures that technological interventions do not merely produce static (either open or of other sort) effects but facilitate continuous emergence, ethical reflection, and social participation.

These three puzzle pieces render my approach both political and ontological, which I term the “political ontology of technology”. Ethics is an integral part of this framework, demonstrating that an ethics of technology – or, in particular, AI – cannot ignore the broader network of co-creators that shape our world. This perspective goes beyond the so-called GELSI framework (governance, ethical, legal, and societal implications) of AI or other technologies. It concerns the very world, ontologically speaking, within which even GELSI conditions emerge.

In the ontic dimension, openness manifests through participatory design, accessible data architectures, localized designs shared openly online, and transparent democratic decision-making systems. But the way such values and principles can be designed or created in regards to AI remains to be researched and evaluated in practice.

In parallel, ontological openness refers to the principles, assumptions, and frameworks through which humans understand technology, knowledge, agency, society, AI, relationality and many more aspects of our world. AI development often embeds implicit ontological assumptions, such as deterministic views of human behavior, algorithmic supremacy, or reductionist models of intelligence. By adopting an ontological lens informed by process philosophy and political ontology of technology, we can interrogate these assumptions, exposing the closures they create and identifying opportunities for openness that foster creativity, reflection, and transformation.

Ultimately, openness is a philosophical and ethical anchor for the development and deployment of AI and of all technology. It integrates design, governance, and societal structures, creating conditions for continuous imaginary emergence, ethical reflection, and democratic participatory engagement. Ethical AI needs to be accounted as inseparable from social, political, and ontological contexts; its evaluation requires attention to the relational, emergent, and multi-scalar character of technology and society. By situating AI within process philosophy, political ontology, and the philosophy of technology – and by centering all of these around the notion of openness, which, as shown, challenges the ontological and political assumptions of contemporary social worlds – we gain conceptual tools to assess both immediate and systemic effects, fostering technologies and institutions that generate ongoing opportunities for ethical, democratic, and creative engagement.

Finally, I would like to express my sincere gratitude to the Center for Hellenic Studies at Harvard University for providing me with the opportunity, funding, and resources to further develop my research. I am deeply grateful to the fellowship committee –Sergio Imparato, Nicolas Prevelakis, and Charalambos Tsekeris – not only for selecting me as a fellow but also for their continuous support, insightful feedback, and many stimulating discussions throughout the fellowship. My special thanks go to Evan Katsarelis for his outstanding organization and coordination of the fellowship’s activities, travel, events, and discussions. His dedication and professionalism greatly contributed to making the experience both intellectually enriching and enjoyable.

It was an unforgettable experience, and I will always be grateful for it.