On the Insecurity of Keystroke-Based AI Authorship Detection: Timing-Forgery Attacks Against Motor-Signal Verification
How much verified proof exists for this claim
One strong evidence source: arxiv
How intriguing or unexplained this claim is
The claim involves an active area of research with ongoing investigations into the security of keystroke-based AI authorship detection, presenting competing theories and notable unknowns about the effectiveness of proposed defenses against specific attacks.
Recent proposals advocate using keystroke timing signals, specifically the coefficient of variation ($δ$) of inter-keystroke intervals, to distinguish human-composed text from AI-generated content. We demonstrate that this class of defenses is insecure against two practical attack classes: the copy-type attack, in which a human transcribes LLM-generated text producing authentic motor signals, and timing-forgery attacks, in which automated agents sample inter-keystroke intervals from empirical...