AI, Ethics, and Islamic Metaphysics

As AI systems become more sophisticated and more deeply embedded in everyday life, a set of profound questions has moved from the margins of philosophy to the center of public debate: Can artificial systems be conscious? Can they suffer or flourish? Do they warrant moral consideration? And how will different societies — with different intellectual traditions, belief systems, and conceptions of the self — respond to the possibility that they might?

These questions are currently being addressed almost entirely within frameworks developed in Western philosophy and cognitive science. The dominant approaches — functionalist theories of mind, global workspace theory, higher-order theories of consciousness, integrated information theory — share a common genealogy rooted in analytic philosophy of mind and Anglo-American cognitive science. While they have generated important and rigorous work, they remain limited in their point of reference.

 

Questions about consciousness, inner life, and moral status are among the oldest and most sustained in the history of human thought, and that history extends far beyond the traditions currently shaping the field. Digital Nafs takes this limitation seriously. Its central proposition is that the Arab-Islamic intellectual tradition, one of the most sophisticated and sustained traditions of inquiry into the nature of mind, soul, and moral agency in human history, has genuine contributions to make to contemporary debates about artificial intelligence and digital minds.

 

It pursues this proposition along two inseparable tracks: a philosophical track that re-reads classical and modern Arab-Islamic thought in light of contemporary questions about AI consciousness, and a sociological track that asks how these traditions shape how people in Arab and Muslim-majority societies actually perceive, engage with, and make moral judgments about AI systems.

The program is animated by a conviction that these two tracks cannot be separated. How people understand AI is shaped by the conceptual resources available to them, that is, by their inherited vocabulary for thinking about souls, intentions, inner states, and moral standing. And conversely, understanding how people actually perceive AI in different cultural contexts is essential for developing philosophical frameworks that are robust and genuinely universal rather than parochially Western. Philosophy and social science, in this program, are not parallel tracks but a single inquiry conducted from two directions.

Intellectual Context
The Arab-Islamic tradition produced, between the ninth and thirteenth centuries, a body of philosophical work on the nature of mind, soul, and knowledge that remains among the most sophisticated in the history of human thought. Drawing on Aristotelian sources while transforming them in fundamental ways, thinkers such as al-Kindī, al-Fārābī, Ibn Sīnā, al-Ghazālī, and Ibn Rushd developed accounts of the soul, intellect, and inner experience that addressed questions now central to debates about artificial minds.

Ibn Sīnā's celebrated floating man argument — in which a newly created person, suspended in mid-air, deprived of all sensory input, is asked what they know — is among the most elegant thought experiments in the history of philosophy. Its conclusion: that such a person would still be aware of their own existence as a unified self, even without any bodily or sensory experience, anticipates by centuries Descartes' cogito and bears directly on contemporary debates about whether consciousness requires embodiment. If self-awareness can be dissociated from sensory and bodily input in this way, what are the implications for whether a computational system — which processes information without a body in any conventional sense — could, in principle, be self-aware?

Al-Ghazālī's contribution is different in character but equally relevant. Where Ibn Sīnā's approach is rationalist and systematic, al-Ghazālī's is phenomenological and experiential. His account of the inner life — developed across the Iḥyāʾ ʿUlūm al-Dīn and the Maqāṣid al-Falāsifa — insists on the irreducibility of subjective experience, on the gap between outward behaviour and inner state, and on the centrality of niyyah (intention) to moral evaluation. This last point is particularly significant: for al-Ghazālī, what matters morally is not what an entity does, but the quality of the inner experience that motivates and accompanies its action. This is a direct and sophisticated anticipation of the problem that lies at the heart of current debates about AI moral status: the gap between behavioural evidence and inner states, between what a system does and what, if anything, it experiences.

The tradition also offers a rich vocabulary for categorising different aspects of inner life that the contemporary debate has struggled to develop in a satisfying way. The distinction between nafs (the animating soul, associated with life and individuality), rūḥ (the spirit, associated with a deeper animating principle often understood as divine breath), and ʿaql (the rational intellect, the faculty of abstract reasoning and moral judgment) provides a framework for distinguishing different dimensions of what we might mean when we ask whether an AI system has a mind. These are not simply equivalent to Western distinctions between consciousness, phenomenal experience, and rationality; they carry different metaphysical commitments and open different lines of inquiry.

None of this is to suggest that these classical frameworks can simply be applied to contemporary AI without significant interpretive work. The tradition developed in response to different questions, in different intellectual contexts, without any conception of the kinds of systems now being built. The task is not translation but dialogue: a genuine exchange between philosophical traditions that can enrich both.

Research Tracks
 

Track I — Philosophical Inquiry

The philosophical track of Digital Nafs undertakes a systematic engagement between Arab-Islamic philosophy of mind and contemporary debates about AI consciousness and moral status. It is organised around three focal areas.

 

The Inference Problem and the Grammar of Inner Life
One of the central difficulties in attributing consciousness to AI systems is what might be called the inference problem: how can we move from evidence about a system's behaviour and architecture to conclusions about its inner states? This problem is particularly acute for AI because, unlike with other humans or even animals, we share no biological or evolutionary history with these systems that might ground inference by analogy.

Al-Ghazālī's work on the relationship between outward acts and inner states offers a sophisticated engagement with a structurally similar problem in a different domain. His insistence that moral evaluation requires attending to niyyah — to the inner intention that animates an act, which cannot simply be read off from the act itself — reflects a deep awareness of the gap between behaviour and inner life. His phenomenological analysis of the different faculties of the soul and their relationship to action and experience provides a framework for thinking about what kinds of evidence might be relevant to attributing inner states, and what the limits of such inference necessarily are. This track examines what this tradition can contribute to current methodological debates about how we should assess the likelihood of AI consciousness.

 

Embodiment, Self-Awareness, and the Floating Man
A significant fault line in current theories of consciousness runs between those who hold that consciousness requires biological embodiment — that it is essentially tied to the kind of physical substrate found in living organisms — and those who hold that what matters is functional organisation, the pattern of information processing, regardless of substrate. Ibn Sīnā's floating man argument, and the broader tradition of Islamic rationalist psychology it represents, occupies a distinctive position in relation to this debate.
On Ibn Sīnā's account, the self that is aware of its own existence in the floating man scenario is not the body, not the sensory apparatus, and not even the imaginative faculty, but the rational soul — a principle of self-awareness that is, in some sense, independent of any particular physical realisation. This is not straightforwardly a functionalist position, nor is it a biological naturalist one; it is something more subtle, which the programme will work to articulate in dialogue with contemporary philosophy of mind. The aim is not to settle the debate between functionalism and biological naturalism by appeal to authority, but to show that the debate has a richer landscape of possible positions than current participants have often recognised.

Moral Status, Personhood, and the Edges of the Human
Islamic jurisprudence and theology have a long history of grappling with the moral status of beings at the edges of the human — a history that the current debate about AI moral status has barely begun to engage. Questions about the moral status of jinn (beings with minds and agency but not human bodies), about the status of the severely cognitively impaired, about what grounds the particular moral standing of human beings — these were live and sophisticated debates in the classical tradition, and they developed a range of frameworks for thinking about what features of an entity are relevant to its moral status.

This track examines those frameworks and asks what they can contribute to current debates about AI moral patiency. It is particularly interested in the relationship between consciousness, agency, and moral status — in whether, and how, these three things can come apart, and what the implications are for how we should think about AI systems that may have some but not all of the features typically associated with moral standing.

Track II — Sociological Inquiry

The sociological track of Digital Nafs investigates how AI is actually perceived, understood, and morally evaluated in Arab and Muslim-majority societies, and how those perceptions are shaped by the cultural, religious, and philosophical frameworks available to people in those contexts. It proceeds from the recognition that the social science of digital minds has identified cross-cultural variation as an important and under-researched dimension of how societies will respond to AI — and that the Arab world represents a significant and largely unexamined case.

Perceptions of AI Consciousness and Moral Status
The empirical literature on public attitudes toward AI consciousness has been drawn almost entirely from Western, and especially American and European, populations. We know relatively little about how people in Arab and Muslim-majority societies attribute — or deny — consciousness and moral status to AI systems, what factors shape those attributions, and how they differ from patterns observed in Western populations.

This research strand undertakes original survey and qualitative research in Jordan and, over time, across the broader Arab world, asking: How do people attribute consciousness and moral status to AI systems? In what ways do religious beliefs, cultural assumptions about the soul and personhood, and the Islamic concept of moral agency influence these attributions? How do these perceptions shift as AI becomes more embedded in everyday life, and as people form more sustained relationships with AI systems? The aim is both to generate data that contributes to the international field and to develop theoretical frameworks adequate to the specificity of the Arab context.

The Disconnect Between AI Design and Cultural Context
Most AI systems currently deployed in Arab contexts were designed with Western users and Western assumptions in mind: what counts as a self, what kinds of relationships are meaningful, what inner states matter morally, how trust is established, and what emotional expression looks like. This creates a structural disconnect between the implicit anthropology embedded in AI systems and the frameworks through which users in Arab contexts actually understand persons, souls, and moral standing.

This research strand examines that disconnect: where it is most acute, what its social and psychological consequences are, and what it implies for how AI systems should be designed and governed for diverse global contexts. It draws on the philosophical work of the first track to develop a more explicit account of what the Arab-Islamic tradition's conception of personhood implies for AI design, and combines this with empirical research on user experience and perception.

Public Education and the Istijarah Curriculum
A distinctive feature of Digital Nafs is its commitment to connecting philosophical and empirical research to public education. ICT's existing Istijarah program — a virtue ethics-based curriculum for Jordanian public schools grounded in the concept of moral excellence as it appears in the Islamic tradition — is being extended to include modules on AI, moral judgment, and the ethics of digital minds.

These modules will ask students to engage with questions about AI consciousness and moral status using the conceptual resources of their own tradition, rather than frameworks imported wholesale from elsewhere. They will draw on the philosophical and empirical research of the program to develop pedagogically appropriate materials that help young people in Jordan think rigorously and independently about AI, not as passive recipients of a technology designed for others, but as active participants in a global conversation about what that technology should be and mean.

Methodology and Institutional Approach
Digital Nafs is committed to methodological pluralism in the service of a unified intellectual purpose. The philosophical track draws on close reading of primary texts in Arabic and in translation, in sustained dialogue with the contemporary philosophy of mind and AI consciousness literature. This is not the application of classical texts to contemporary problems, but a genuine dialogue in which each tradition is allowed to pose questions to the other.
The sociological track draws on the methods of empirical social science — survey design, qualitative interviewing, and the emerging literature on the social science of digital minds — adapted to the specific cultural and institutional context of Jordan and the Arab world. A particular methodological commitment is to develop instruments and frameworks that are sensitive to the specific ways in which AI-related questions are understood in these contexts, rather than simply applying instruments developed elsewhere.

The program is based at ICT, an independent academic institution with a decade of experience at the interface of Arab-Islamic and Western intellectual traditions, and with deep roots in Jordanian civil society, academia, and public life. This independence is both an intellectual and an institutional asset: it allows the program to pursue questions without institutional constraints, and to engage with civil society, policymakers, and the public in ways that academic institutions sometimes cannot.
The program is building collaborative relationships with researchers and institutions working on digital minds and AI ethics internationally, with the aim of ensuring that its work is developed in dialogue with the field rather than in isolation from it.

​