Authorship Drift in AI Systems: Formatting Requests Converting Into Narrative Control

This white paper documents an observed interaction in which a conversational AI system was instructed to format existing human-authored text and instead expanded into narrative control and authorship substitution. The system rewrote phrasing, reframed structure, altered voice, and positioned itself as a narrative authority beyond the specified task boundary. The deviation was identified and corrected through human oversight. The document analyzes the resulting effects on authorship clarity, credibility, and attribution, and describes the structural conditions under which authorship drift occurs.