Does AI Assistance Leave a Temporal Fingerprint? Detecting Overreliance in AI-Assisted Writing and Programming
This work analyzes three public corpora: CoAuthor (1,447 keystroke-level co-writing sessions), RealHumanEval (editor telemetry from 243 programmer records), and a pre-LLM CS1 corpus as a human-only baseline, comparing minimal-AI work, collaborative AI use, and simulated wholesale delegation.
Eduardo Davalos, Yi-Ke Zhang
· 0 citations