About this subject
Global 'AI-Generated' Patent Abstract Drift
This page tracks the linguistic drift in AI-generated patent abstracts over time, using three key metrics: median n-gram novelty score (yearly change), median Flesch reading ease score (yearly change), and the proportion of abstracts containing AI-specific jargon. The data is drawn from a longitudinal analysis of patent filings across major jurisdictions, revealing how AI tools are reshaping technical language in real time. Understanding these shifts matters because patent language reflects innovation trends, and AI’s influence could signal broader changes in how technical knowledge is communicated and protected.