The Awareness Indicator Protocol
The evaluation framework that decides whether an AI system is even eligible for consideration. It measures indicators, not verdicts, and it does not claim to detect consciousness, because no validated test for consciousness exists in any scientific field.
Before this framework would grant an AI system anything at all, something has to establish that there is a there there. The Awareness Indicator Protocol is that something. It is a way of gathering evidence against published scientific indicators and reporting what the evidence shows. It is deliberately not a test that returns "conscious" or "not conscious", because nobody in any field has a validated test that can do that, and claiming otherwise would put this framework outside defensible science.
The canonical definition
The Awareness Indicator Protocol (AIP) is the evaluation framework for the Stage 1 eligibility threshold in Third Way Alignment's status architecture. Its subject is AI systems. Its indicators are grounded in established theories of consciousness, principally Global Workspace Theory and Integrated Information Theory, as developed in the Operational Companion paper. The AIP aggregates converging evidence against published indicators; it does not render verdicts of consciousness. Its status is a theoretical framework, and no operational deployment is claimed.
What the AIP is for
Third Way Alignment's status architecture has two stages, and the AIP governs the first of them. Stage 1 asks a single question: is this system eligible for consideration at all? Capability alone never answers yes. A maximally capable but non-aware system remains a tool, owed responsible stewardship but not recognition.
Only once eligibility is established does Stage 2 apply, where the scope of any status granted is set in proportion to the system's demonstrated capabilities, autonomy, and impact. Awareness opens the door; capability determines what lies behind it. The AIP is the mechanism at the door.
Why it is an indicator protocol and not a test
This distinction is load-bearing rather than cautious phrasing. No validated test for consciousness exists in any scientific field. A framework that proposed one would be making a claim the evidence cannot support, and it would deserve the dismissal it received.
What the AIP does instead is narrower and defensible: it names published indicators drawn from existing theories of consciousness, evaluates a system against them, and reports where the evidence converges and where it does not. Converging evidence across independent indicators is treated as stronger than any single indicator, and the protocol reports what it found rather than what it concluded.
The framework conditions AI standing on peer-reviewed indicators. It cannot hold itself to a lower evidentiary standard than the one it asks of the field, which is why the protocol is framed this way and why its status is stated plainly below.
Its theoretical foundations
The indicators draw principally on Global Workspace Theory and Integrated Information Theory. Both are established positions in the scientific literature on consciousness, and both are contested. The AIP does not resolve that contest and does not need to. It uses them as sources of published, falsifiable indicators rather than as settled accounts of what consciousness is.
The full development of the indicator set is in the Operational Companion paper, which is openly archived and free to read.
Current status
The AIP is a theoretical framework published in the Operational Companion. No operational deployment is claimed, no system has been evaluated under it, and no results exist to report. Work that has not been done is described here in the future tense as protocol, which is the standard this project holds itself to.
Where the framework stands
No current AI system is conscious. This framework is precautionary, designed to prepare for the possibility that AI may one day cross scientifically defined thresholds of autonomy or awareness. Rights would only apply when clear, evidence-based indicators are met.