works
Samuel Shadrach The case for studying stylometric deanonymisation as surveillance tech online The article argues that the development of artificial intelligence (AI)-based stylometric deanonymisation tools presents a serious threat to the possibility of achieving an unsurveillable society. Stylometrics, the study of linguistic style, allows for the identification of individuals based on their writing patterns, which are unique and persistent even when attempting to obscure one’s identity. While current stylometric tools have limited success, the author advocates for further research into AI-based stylometrics as a way to gauge its potential for deanonymisation. The author acknowledges the potential infohazard associated with this research, but believes that the knowledge gained, regardless of its outcome, is essential for developing effective privacy-preserving technologies and for understanding the limitations of existing anonymity-enhancing tools. – AI-generated abstract.

The case for studying stylometric deanonymisation as surveillance tech

Samuel Shadrach

Effective Altruism Forum, June 19, 2022

Abstract

The article argues that the development of artificial intelligence (AI)-based stylometric deanonymisation tools presents a serious threat to the possibility of achieving an unsurveillable society. Stylometrics, the study of linguistic style, allows for the identification of individuals based on their writing patterns, which are unique and persistent even when attempting to obscure one’s identity. While current stylometric tools have limited success, the author advocates for further research into AI-based stylometrics as a way to gauge its potential for deanonymisation. The author acknowledges the potential infohazard associated with this research, but believes that the knowledge gained, regardless of its outcome, is essential for developing effective privacy-preserving technologies and for understanding the limitations of existing anonymity-enhancing tools. – AI-generated abstract.

PDF

First page of PDF