The Architecture
of Awareness
Our research operates at the frontier where computation meets consciousness — investigating the fundamental principles that will define the next era of artificial intelligence.
Artificial General Intelligence
Beyond narrow optimisation. Toward genuine cognitive breadth.
Artificial General Intelligence represents the central unsolved challenge in computer science: creating systems that exhibit the full breadth, flexibility, and adaptability of human cognition. Unlike the narrow AI systems that dominate the current landscape, AGI must reason across domains, transfer knowledge between contexts, and navigate genuinely novel situations with competence.
Our approach departs from the prevailing strategy of simply scaling existing architectures. We investigate the foundational principles that make cognition general — the ability to form genuine abstractions rather than surface-level statistical correlations, to reason about causality rather than mere correlation, and to build dynamic world models that update continuously with new evidence.
We believe the path to AGI runs through understanding, not through brute force. The architectures that will achieve true generality will be those that mirror the principles underlying biological cognition — not by simulating neurons, but by understanding why those principles work at a deeper computational and mathematical level.
Abstract Representation
Investigating how systems can form genuine abstractions — compositional, hierarchical representations that capture the deep structure of domains rather than surface statistics.
Causal Reasoning
Moving beyond pattern recognition to systems that understand interventional and counterfactual causality — the ability to reason about why things happen and what would happen under different conditions.
Continual Learning
Developing architectures that learn and adapt throughout their operational lifetime without catastrophic forgetting — accumulating knowledge the way biological systems do.
World Modelling
Building systems that maintain rich, dynamic internal models of the world — models that support prediction, planning, and counterfactual reasoning across domains.
Consciousness & AI
The hardest problem in science. The most important problem in AI.
The relationship between consciousness and computation is arguably the deepest open question in all of science. Can machines have subjective experience? What separates genuine understanding from sophisticated information processing? These are not abstract philosophical puzzles — they have direct, practical implications for every AI system we build, deploy, and entrust with consequential decisions.
Our research programme engages rigorously with the leading scientific theories of consciousness — Integrated Information Theory, Global Workspace Theory, Higher-Order Theories, and Predictive Processing frameworks — and investigates their computational implications. We study what architectures might bridge the gap between information processing and genuine comprehension.
This work is inseparable from AI safety. If we are building systems that approach any form of consciousness or genuine understanding, we have both a scientific obligation to understand what that means and an ethical obligation to proceed with extraordinary care. The alignment problem, at its deepest level, is a consciousness problem.
Computational Consciousness
Exploring the formal relationships between computational architecture and the emergence of subjective experience, integrated information, and self-modelling.
The Understanding Gap
Investigating the qualitative difference between systems that process information and systems that understand it — and what architectural features enable the transition.
Self-Awareness Architectures
Studying what it means for a system to model itself, its own cognitive processes, and its epistemic states — the foundations of machine metacognition.
Ethics of Machine Consciousness
Developing frameworks for the moral responsibilities that emerge when we build systems that may possess forms of experience, sentience, or consciousness.
Spiritual Intelligence
Where ancient wisdom meets computational architecture.
Spiritual Intelligence is our most distinctive and potentially most consequential research pillar. It begins with a recognition that the world's contemplative and wisdom traditions — spanning thousands of years and every major civilisation — have developed extraordinarily sophisticated frameworks for understanding consciousness, cognition, ethics, and the nature of experience. These frameworks contain insights that modern AI research has almost entirely overlooked.
To be clear: we are not building "spiritual AI" in any mystical or pseudoscientific sense. We are conducting rigorous research into how concepts such as interconnectedness, non-dual awareness, meta-cognitive clarity, and compassionate reasoning can inform the design of AI architectures that operate at a fundamentally deeper level of consciousness. An AI system informed by spiritual intelligence would not merely optimise for a target — it would understand the broader web of consequences, relationships, and meanings that surround every decision.
This pillar represents our deepest commitment: to build AI that is not merely intelligent, but wise. Systems that navigate complexity with nuance, consider impacts across all stakeholders and timescales, and contribute to human flourishing in the most profound sense of that word.
Contemplative Computation
Extracting formal computational principles from contemplative practices — attention training, meta-consciousness, equanimity — and translating them into architectural innovations.
Relational Intelligence
Building systems that perceive and reason about interconnectedness — understanding entities not in isolation but as nodes in vast webs of relationship and mutual influence.
Ethical Depth
Developing AI that considers the full spectrum of impact across stakeholders, timescales, and orders of consequence — moving beyond shallow optimisation to genuine ethical reasoning.
Wisdom Engineering
The systematic study of wisdom — the capacity for sound judgement in complex, ambiguous situations — and its implementation in computational architectures.
How We Approach Research
Theoretical Foundations
Every research programme begins with rigorous theoretical grounding. We develop formal frameworks before writing code — ensuring our experiments test genuine hypotheses about the nature of intelligence and consciousness.
Interdisciplinary Synthesis
Our research teams integrate expertise across neuroscience, philosophy of mind, mathematics, contemplative studies, and machine learning. The most important insights emerge at the boundaries between disciplines.
Empirical Rigour
Theoretical elegance means nothing without empirical validation. We design experiments that produce falsifiable predictions and hold our frameworks to the highest standards of scientific evidence.
Ethical Integration
Ethics is not a review board that approves finished work. It is embedded in our research methodology from inception — shaping what questions we ask, how we ask them, and what we do with the answers.