Philosophers Without Philosophy: Conceptual Boundary Integrity in AI-Consciousness Discourse

Philosophy does not fail when it becomes difficult; it fails when it becomes non-examinable. Original tool CBIT: a test for definition, method, evidence, and metaphor control. Published cases Dennett-style imitation, LLM consciousness, and quantum-consciousness overextension. Aim Defend interdisciplinary dialogue from method substitution and false depth. Boundary is not censorship; boundary is the condition that makes dialogue intelligible. Abstract This paper develops a diagnostic account of boundary failure in contemporary debates about philosophy, consciousness, religion, mysticism, neuroscience, physics, and artificial intelligence. Its central claim is not that difficult philosophy is a defect, but that discourse becomes philosophically weak when difficulty is produced by conceptual evasion: when concepts are undefined, methods are silently substituted, evidence is borrowed from the wrong domain, and metaphor is treated as proof. The paper names this pattern new sophistry when discourse becomes non-examinable, and it uses digital gnosticism as a critical metaphor for the premature elevation of AI systems into hidden beings, digital souls, or inward consciousnesses before adequate conceptual and empirical bridges are supplied. Its central contribution is the Conceptual Boundary Integrity Test (CBIT), a four-part diagnostic framework requiring claims to define the concept, specify the method, ask for domain-appropriate evidence, and control the metaphor. To strengthen the framework beyond general critique, the paper applies CBIT to three published debates: philosophical imitation in a language model of Daniel Dennett, claims about large language model consciousness and AI welfare, and quantum-consciousness arguments that are often extended into quantum-spiritual rhetoric. The argument does not reject difficult philosophy, religious meaning, mystical experience, scientific explanation, or interdisciplinary dialogue. It rejects only the uncontrolled substitution of one domain's language for another domain's proof. By distinguishing dialogue from confusion, the paper offers a practical framework for evaluating claims about AI consciousness, machine suffering, neural correlates, quantum-spiritual rhetoric, and the philosophical interpretation of technology. Highlights • Introduces CBIT as a compact test for conceptual and methodological accountability. • Applies the test to three published debates rather than relying only on general critique. • Distinguishes interdisciplinary dialogue from unlicensed substitution between domains. • Clarifies how metaphor can guide inquiry without becoming proof.