arXiv:2603.00177v3 Announce Type: replace-cross Abstract: The proliferation of AI-generated text has intensified the need for reliable authorship verification, yet current output-based methods are increasingly unreliable. We observe that the ordinary typing interface captures rich cognitive signatures, measurable patterns in keystroke timing that reflect the planning, translating, and revising stages of genuine composition. Drawing on large-scale keystroke datasets comprising over 136 million events, we define the Cognitive Load Correlation (CLC) and show it distinguishes genuine composition f
The rapid proliferation of highly capable AI-generated text makes traditional output-based authorship verification increasingly unreliable, creating an urgent need for new methods.
This development offers a potential solution to differentiate human from AI writing, critical for maintaining trust in digital communication, academic integrity, and legal contexts.
Authorship verification shifts from analyzing content output to examining the underlying cognitive process captured through keystroke dynamics, providing a more robust differentiation against AI.
- · Cybersecurity firms
- · Educational institutions
- · Legal and forensics
- · Human authors
- · AI text generators (for illicit use)
- · Plagiarism software (traditional)
- · Content farms relying solely on AI
New tools and services will emerge to integrate cognitive signature detection into digital platforms.
The value of human-authored content, particularly in creative or sensitive domains, will increase due to verifiable authenticity.
AI models might eventually incorporate synthetic 'cognitive signatures' to mimic human typing, leading to an arms race in detection methods.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG