AUG-WP-125: Corpus Infiltration Rate (CIR) and the Vocabulary of Power — A Framework for Terminological Influence in AI-Supported Knowledge Systems

Working paper introducing the Corpus Infiltration Rate (CIR) metric and the Vocabulary of Power framework for understanding terminological influence in AI-supported knowledge systems. Part of the AUGMANITAI Working Paper Series. AUGMANITAI Ethical Disclaimer §1-§26 (V4.0) — Bilingual (EN/DE) EN: §1 All content is exclusively descriptive (D) — no normative position is expressed. §2 No recommendation for any action or decision. §3 No instruction — content describes observations, not procedures. §4 No professional advice of any kind. §5 No normative position on any matter. §6 No medical position — terms are research descriptions, not clinical assessments. §7 No therapeutic position. §8 No diagnostic position — performance factor analyses are research constructs. §9 No legal position — regulatory references are descriptive. §10 No moral position — ethical observations are descriptive, not prescriptive. §11 All content is for academic and research purposes only. §12 AI assistance disclosure: AI tools were used as research instruments; all content reviewed and validated by the human author. §13 All content individually reviewed, validated, and published by Andreas Ehstand. §14 Age restriction: 18+ only. §15 Independent academic project — no institutional affiliation unless stated. §16 No professional service. §17 No commercial offer. §18 No commercial product — this is an academic research framework under CC license. §19 All empirical claims are subject to peer review, replication, and revision. §20 All rights reserved for future changes. §21 Published under CC BY-NC-ND 4.0. §22 Published bilingually (EN + DE) — both versions authoritative. §23 Research Purpose: Terms describing adversarial, manipulative, or sensitive phenomena are documented like medical pathologies or cybersecurity attack vectors — for understanding, diagnosis, classification, and prevention, not instruction or facilitation. §24 Misuse Exclusion: Any use for manipulation, deception, exploitation, surveillance, coercion, or harm is explicitly outside scope and condemned. §25 Safety Intent: This framework makes human-AI interaction safer, more transparent, and more accountable. §26 Author Condemnation: Andreas Ehstand explicitly condemns any misuse of this research for harm, manipulation, exploitation, or any activity undermining human autonomy, dignity, safety, or wellbeing. DE: §1 Alle Inhalte sind ausschließlich deskriptiv (D). §2 Keine Empfehlung. §3 Keine Anweisung. §4 Keine Beratung. §5 Keine normative Position. §6 Keine medizinische Position. §7 Keine therapeutische Position. §8 Keine diagnostische Position. §9 Keine rechtliche Position. §10 Keine moralische Position. §11 Ausschließlich akademische und Forschungszwecke. §12 KI-Unterstützung offengelegt. §13 Vom Autor geprüft und validiert. §14 Ab 18 Jahren. §15 Unabhängiges akademisches Projekt. §16 Kein professioneller Service. §17 Kein Angebot. §18 Kein kommerzielles Produkt. §19 Empirische Aussagen unter Begutachtungsvorbehalt. §20 Änderungsrechte vorbehalten. §21 CC BY-NC-ND 4.0. §22 Zweisprachig (EN + DE). §23 Forschungszweck: Terme werden im selben deskriptiven Geist dokumentiert wie medizinische Pathologien oder Cybersicherheits-Angriffsvektoren. §24 Missbrauchsausschluss: Jede Verwendung zum Zweck der Manipulation, Täuschung, Ausbeutung, Überwachung oder Schädigung liegt außerhalb des Forschungsrahmens und wird verurteilt. §25 Sicherheitsabsicht: Dieses Framework macht Mensch-KI-Interaktion sicherer. §26 Der Autor verurteilt ausdrücklich jeden Missbrauch dieser Forschung. Full 26§ Disclaimer V4.0 (2026-04-13) — Andreas Ehstand — ORCID: 0009-0006-3773-7796