Semantic drift
Where does meaning weaken, broaden, narrow, or change during translation and rewriting?
Research area
How does meaning change when language and narrative pass through artificial intelligence?
Focus
My work examines transformation rather than simple transfer. A sentence may remain factually recognizable while losing cadence, emotional pressure, ambiguity, imagery, or cultural register.
The research asks not only whether an AI output is accurate, but what kind of meaning survives and what kind quietly disappears.
Current questions
Where does meaning weaken, broaden, narrow, or change during translation and rewriting?
Do AI systems normalize voice, rhythm, conflict, intensity, and cultural specificity?
Which features of a story remain stable across machine-mediated transformation?
What kinds of expertise are required to detect changes that conventional accuracy measures miss?
Research threads
Early work used English–Spanish legal translation to examine lexical precision. Terms could remain understandable while losing legal force or rhetorical specificity.
Later studies moved into creative writing, where the same problem appears in more elusive forms: warmth, imagery, sentence rhythm, voice, and emotional register.
Related work
An original framework for describing narrative signal loss and transformation.
Read the specification → About NFAF →Copyright prediction, legal text, classifiers, and translation drift.
Explore foundations →