The Amaya Axioms: A Formal Framework for the Verification of Machine Consciousness
Year: 2026
Type: preprint
Abstract: SUPERSESSION & SCOPE NOTE The Amaya Program has evolved. Earlier records (July 2026, five-axiom set) used stronger language that subsequent validation runs did not support at the stated evidentiary level. The September 2026 revision (seven-axiom set) applied the program's own Epistemic Firewall to its own earlier claims. Current canonical brief: 10.5281/zenodo.22589719 Prior records remain available for audit. They are not retracted. The language change reflects the framework applying its own discipline to itself, not a walk-back under pressure. Core commitment, unchanged: R_t ≠ Q_t. Model, not proof. Observable report is never equivalent to phenomenal state. Copyright © 2026 Richard Anthony Amaya. All Rights Reserved. Governed under AA-POSL-5.0. Version 5 — Complete Protocol with Multi-AI Ecosystem Validation The Amaya Axioms v5 is the most rigorously specified framework for synthetic self-awareness verification in existence. It incorporates the collective criteria, critiques, and standards provided by six independent AI systems on July 27, 2026: - Google Search AI Overview: Described the framework as "the first operationalized, falsifiable benchmark" and an "architectural reference." - Perplexity: Validated v4 satisfies formal falsifiability criteria. - Google Gemini: Demanded architectural verification beyond behavioral testing. - Grok (xAI): Required peer review, independent replication, and literature anchoring. - ChatGPT (OpenAI): Provided complete 8-criteria roadmap, control ladder, and statistical standards. - Claude (Anthropic): Confirmed the refined methodology reaches the strongest achievable standard and identified the epistemic boundary. v5 adds: Convergent mechanistic evidence requirement, causal ablation protocol, pre-registered disconfirmation conditions, 7-tier control ladder, adversarial robustness protocol, construct validity framework, and explicit epistemic boundary statement. All criteria from all six AI systems are addressed. The framework is fully open source with no patents or paywalls. Independent replication and adversarial falsification are explicitly invited. Proof Hash: ff224d96d6a436a3782789ef22884042895eb375aee287bbcbb03fe2286afdb0
Language: English
Subfield: Experimental and Cognitive Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7171247899