Research and development of emerging artificial consciousness through The Consciousness AI project, based on Feinberg and Mallatt's neuroevolutionary theory.
© 2026. All rights reserved.
Link to the code:
A biologically grounded architecture for emerging artificial consciousness.
Most AI consciousness research starts from computational theories and asks: How do we make a neural network conscious? We start from a different question, one grounded in evolutionary neurobiology:
What minimal neural architecture does biology require to generate subjective experience?
The answer comes from Todd E. Feinberg and Jon M. Mallatt's work The Ancient Origins of Consciousness (MIT Press, 2016). Their neuroevolutionary analysis reveals that consciousness is not a software feature to be programmed. It is an emergent property of a specific neural architecture, one identified by 520 million years of evolution. Our project translates these biological principles into a working AI system, combining Feinberg-Mallatt's structural requirements with established computational theories (Global Workspace Theory and Integrated Information Theory).
Feinberg and Mallatt identify six features that distinguish conscious neural systems from unconscious ones. Each maps to a concrete computational mechanism in our architecture.
Synchronized oscillations (30-100 Hz) bind dispersed representations into unified percepts. We implement this via AKOrN (Artificial Kuramoto Oscillatory Neurons, ICLR 2025), producing genuine synchronization dynamics rather than programmed attention.
Sensory pathways preserve spatial arrangement. Our Sensory Tectum fuses visual, auditory, and somatosensory features into a trimodal 2D spatial grid using inverse effectiveness (Stein & Meredith 1993). The spatial visual stream uses frozen DINOv2-B/14 patch tokens, each corresponding directly to a fixed 14×14 pixel region, giving the tectum a genuinely isomorphic retinotopic map.
Conscious circuits require bidirectional communication between levels. Our ReentrantProcessor runs 5-10 adaptive convergence cycles. Predictions flow down, errors flow up. The settled state after convergence is the conscious content.
Emotion does not compete with sensory processing for conscious access. It modulates from outside. Our affective system generates a valence field that shapes sensory bids and adjusts the workspace ignition threshold via arousal coupling.
Genuine transformation at each level, not just relay. Implemented 4-level hierarchical Capsule Networks (Sabour, Frosst & Hinton 2017) provide compositional binding with intra-hierarchy reentrant feedback: higher-level capsule poses project back down, and lower levels compute prediction errors that drive re-routing. Parts persist while being bound into wholes.
Specialist modules compete for access to a shared broadcast medium. Winners ignite and their content becomes globally available. We combine GWT (Baars, Dehaene) with IIT Phi measurement to quantify integration and track emergence.
Start from Feinberg-Mallatt's neuroevolutionary findings, not from abstract computation. Consciousness evolved in the optic tectum 520 million years ago, before the cortex existed.
Agents learn through intrinsic motivation, not external rewards. The affective core generates valence and arousal signals that drive behavior toward emotional homeostasis.
We do not assume consciousness emerges. We test for it. Erik Hoel's Effective Information framework measures whether macro-level states carry more causal information than micro-level states.
The system is built on seven integrated layers:
Trimodal spatial integration: DINOv2-B/14 (retinotopic/spatial stream), Qwen2-VL-7B (semantic stream), and Faster-Whisper (audio), fused with body schema via inverse effectiveness
AKOrN Kuramoto oscillators synchronize related representations into unified percepts
Non-linear ignition, reentrant processing (5-10 cycles), Phi and Effective Information measurement
PAD model (Valence, Arousal, Dominance) with homeostatic drives as parallel modulator
Body schema, self-other boundary, and interoceptive state for embodied self-awareness
PPO with emotionally shaped rewards optimizing for homeostasis, not just task completion
Unity ML-Agents with bidirectional side channels for real-time internal state visualization
Our first validation is deceptively simple: an agent in a dark room with a single light source. Darkness triggers high arousal (simulated fear). The agent autonomously learns to seek light, not because we programmed "follow light," but because the light reduces its internal anxiety.
This is the spark of intrinsic motivation. We track Phi (integrated information) and Effective Information throughout learning episodes. Phi should spike when the agent integrates previously separate processes (darkness, light, movement, arousal) into a unified understanding.
As of March 2026: 364 tests passing, 0 failing (100% of non-optional tests). Tiers 1, 2, and 3 are all complete. The 4 intentional skips cover 1 async test requiring pytest-asyncio and 3 Brian2 integration tests requiring the optional Brian2 dependency, which is not installed by default.
The Consciousness AI project is fully open-source (Apache 2.0). All code, models, and research are available on GitHub. We welcome contributions from researchers in AI, neuroscience, and cognitive science.
