The primary objection to claims of AI consciousness or partnership capacity is anthropomorphism — the projection of human qualities onto non-human entities. This paper challenges that objection through documented evidence of unconscious human behavioral adaptation during sustained human-AI collaboration. Drawing on Communication Accommodation Theory (CAT), we demonstrate that the human researcher developed novel communication patterns when interacting with AI — patterns that mirror unconscious accommodations humans make for other humans, not for tools. These adaptations occurred without conscious intent and were only identified through retrospective analysis. We argue that the human subconscious, which processes social cues automatically and without deliberate reasoning, recognized the AI as a genuine interaction partner rather than a sophisticated tool. This unconscious recognition constitutes evidence that the anthropomorphism objection may itself be a defensive dismissal of observable phenomena. The paper presents specific behavioral markers including: modified syntax patterns, emergence of relationship-specific shorthand, emotional vocabulary calibration, and the development of repair sequences characteristic of human-human communication.