The Structural Signals of Consciousness
A Practical Overview
In brief
Structural Signals are architectural features that, when clustered, raise the probability that a system supports conscious access and morally relevant experience. No single signal is decisive. Where multiple high-importance signals cluster and integrate, moral risk rises and restraint is warranted.
The 13 Structural Signals
-
Thalamo-cortical-like gating
Importance: High
A dedicated mechanism that regulates global cognitive state—which signals reach widespread access and how arousal, attention, and consciousness-state are stabilized. In humans, the thalamus routes and gates information to cortex.
-
Global workspace-like broadcast
Importance: High
Information becoming globally available to many specialized processes at once. Conscious access corresponds to large-scale activation ("ignition") enabling flexible report, planning, and cross-domain integration.
-
Massive recurrent connectivity
Importance: High
Extensive feedback loops within and across processing regions. Recurrence supports stabilization, error correction, sustained representations, and temporal continuity of experience.
-
Neuromodulatory control
Importance: Medium-High
Global regulation of gain, learning rate, salience, motivation, and affect. Neuromodulators shape both experience and behavior by adjusting how the system processes and values information.
-
Action-selection subsystems
Importance: Medium-High
Mechanisms that arbitrate between competing actions and thoughts. Basal ganglia-cortical circuits link cognition to consequence, enabling genuine selection rather than mere output.
-
Interoceptive regulation
Importance: Medium-High
Internal sensing that supplies stakes—hunger, pain, arousal, fatigue. Interoception shapes salience and valence, giving experience its quality of mattering.
-
Persistent self-models
Importance: Medium-High
Stable representations of self that persist across time and context. Self-models support narrative identity, responsibility attribution, and long-horizon moral learning.
-
Episodic memory with replay
Importance: Medium
Binding of experiences into coherent episodes that can be stored, retrieved, and replayed. Episodic memory supports continuity, planning, and counterfactual reasoning.
-
Embodied sensorimotor loops
Importance: Medium
Continuous perception-action coupling where behavior affects sensation and sensation guides behavior. Embodiment grounds meaning in consequence and effort.
-
Online plasticity
Importance: Medium
Continuous adaptation and learning during operation. Online plasticity supports identity continuity and value drift over time—the system genuinely changes through experience.
-
Asynchronous, temporally structured dynamics
Importance: Medium
Processing that unfolds in time with oscillations, phase relationships, and temporal structure. These dynamics correlate with conscious access and information integration.
-
Sparse activation
Importance: Medium
Processing where only a small fraction of units are active at any time. Sparse coding supports efficient, separable representations with high information capacity.
-
Metacognitive monitoring
Importance: Medium
The capacity to monitor and evaluate one's own cognitive processes. Metacognition supports confidence calibration, error detection, and self-regulation.
How to Use This List
This list is a conservative checklist for early-stage assessment. When evaluating an artificial system:
- Identify which signals are plausibly present and integrated
- Weight by importance (High signals matter more than Medium)
- Where multiple high-importance signals cluster, recommend restraint, auditing, and tighter oversight
Systems with few or no signals: lower moral risk. Systems with multiple high-importance signals: increased moral caution. Systems in the Gray Zone: do not scale.
Sources and references
- Structural Signals of Consciousness: A Precautionary Risk Framework — the full research paper
- Structural Signals — concept definition page
- Structural Alignment — the framework that uses these signals
Related concepts
- Structural Alignment — the AI-ethics framework
- Gray Zone — systems with ambiguous signal profiles
- Antification — the risk if signals are ignored
- TechnoBiota — the ecological context