RECORD
Silencing the Risk, Not the Whistle: A Semi-automated Text Sanitization Tool for Mitigating the Risk of Whistleblower Re-Identification
Computers and Society · Submitted 2 May 2024
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim Our method can significantly reduce AA accuracy from 98.81% to 31.22%, while preserving up to 73.1% of the original content's semantics.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Whistleblowing is essential for ensuring transparency and accountability in both public and private sectors. However, (potential) whistleblowers often fear or face retaliation, even when reporting anonymously. The specific content of their disclosures and their distinct writing style may re-identify them as the source. Legal measures, such as the EU WBD, are limited in their scope and effectiveness. Therefore, computational methods to prevent re-identification are important complementary tools for encouraging whistleblowers to come forward. However, current text sanitization tools follow a one-size-fits-all approach and take an overly limited view of anonymity. They aim to mitigate identification risk by replacing typical high-risk words (such as person names and other NE labels) and combinations thereof with placeholders. Such an approach, however, is inadequate for the whistleblowing scenario since it neglects further re-identification potential in textual features, including writing style. Therefore, we propose, implement, and evaluate a novel classification and mitigation strategy for rewriting texts that involves the whistleblower in the assessment of the risk and utility. Our prototypical tool semi-automatically evaluates risk at the word/term level and applies risk-adapted anonymization techniques to produce a grammatically disjointed yet appropriately sanitized text. We then use a LLM that we fine-tuned for paraphrasing to render this text coherent and style-neutral. We evaluate our tool's effectiveness using court cases from the ECHR and excerpts from a real-world whistleblower testimony and measure the protection against authorship attribution (AA) attacks and utility loss statistically using the popular IMDb62 movie reviews dataset. Our method can significantly reduce AA accuracy from 98.81% to 31.22%, while preserving up to 73.1% of the original content's semantics.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Whistleblowing is essential for ensuring transparency and accountability in both public and private sectors. | background 1.00 | 0.03 | 0.02 | asserted 1.00 |
| s1 | However, (potential) whistleblowers often fear or face retaliation, even when reporting anonymously. | background 1.00 | 0.04 | 0.02 | asserted 0.66 |
| s2 | The specific content of their disclosures and their distinct writing style may re-identify them as the source. | background 1.00 | 0.04 | 0.02 | hedged possibility 0.94 |
| s3 | Legal measures, such as the EU WBD, are limited in their scope and effectiveness. | background 0.78 | 0.05 | 0.02 | asserted 1.00 |
| s4 | Therefore, computational methods to prevent re-identification are important complementary tools for encouraging whistleblowers to come forward. | background 0.79 | 0.05 | 0.10 | asserted 0.98 |
| s5 | However, current text sanitization tools follow a one-size-fits-all approach and take an overly limited view of anonymity. | gap 0.85 | 0.05 | 0.04 | asserted 1.00 |
| s6 | They aim to mitigate identification risk by replacing typical high-risk words (such as person names and other NE labels) and combinations thereof with placeholders. | background 0.96 | 0.05 | 0.02 | asserted 1.00 |
| s7 | Such an approach, however, is inadequate for the whistleblowing scenario since it neglects further re-identification potential in textual features, including writing style. | gap 0.91 | 0.09 | 0.07 | asserted 1.00 |
| s8 | Therefore, we propose, implement, and evaluate a novel classification and mitigation strategy for rewriting texts that involves the whistleblower in the assessment of the risk and utility. | contribution 1.00 | 0.93 | 0.23 | asserted 0.99 |
| s9 | Our prototypical tool semi-automatically evaluates risk at the word/term level and applies risk-adapted anonymization techniques to produce a grammatically disjointed yet appropriately sanitized text. | method 0.94 | 0.92 | 0.09 | asserted 1.00 |
| s10 | We then use a LLM that we fine-tuned for paraphrasing to render this text coherent and style-neutral. | method 1.00 | 0.92 | 0.04 | asserted 1.00 |
| s11 | We evaluate our tool's effectiveness using court cases from the ECHR and excerpts from a real-world whistleblower testimony and measure the protection against authorship attribution (AA) attacks and utility loss statistically using the popular IMDb62 movie reviews dataset. | method 0.99 | 0.96 | 0.05 | asserted 0.99 |
| s12 | Our method can significantly reduce AA accuracy from 98.81% to 31.22%, while preserving up to 73.1% of the original content's semantics. | result 1.00 | 0.97 | 0.78 | asserted 0.86 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Whistleblowing is essential for ensuring transparency and accountability in both public and private sectors. However, (potential) whistleblowers often fear or face retaliation, even when reporting anonymously. The specific content of their disclosures and their distinct writing style may re-identify them as the source. Legal measures, such as the EU WBD, are limited in their scope and effectiveness. Therefore, computational methods to prevent re-identification are important complementary tools for encouraging whistleblowers to come forward. However, current text sanitization tools follow a one-size-fits-all approach and take an overly limited view of anonymity. They aim to mitigate identification risk by replacing typical high-risk words (such as person names and other NE labels) and combinations thereof with placeholders. Such an approach, however, is inadequate for the whistleblowing scenario since it neglects further re-identification potential in textual features, including writing style. Therefore, we propose, implement, and evaluate a novel classification and mitigation strategy for rewriting texts that involves the whistleblower in the assessment of the risk and utility. Our prototypical tool semi-automatically evaluates risk at the word/term level and applies risk-adapted anonymization techniques to produce a grammatically disjointed yet appropriately sanitized text. We then use a LLM that we fine-tuned for paraphrasing to render this text coherent and style-neutral. We evaluate our tool's effectiveness using court cases from the ECHR and excerpts from a real-world whistleblower testimony and measure the protection against authorship attribution (AA) attacks and utility loss statistically using the popular IMDb62 movie reviews dataset. Our method can significantly reduce AA accuracy from 98.81% to 31.22%, while preserving up to 73.1% of the original content's semantics.
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
What it cites · 7 sentences without a confident label
Uses the method or tool of 2 papers
-
We assess AAA using the Valla software package [ref] , specifically its two most effective models: one based on character n-grams and the other on BERT.
On the State of the Art in Authorship Attribution and Authorship Verification · Data, language models, and settings · 1.00
-
The first measure uses RoBERTa [ref] to assess how surprising an entity is in its context, assuming that more unique entities carry higher privacy risks.
RoBERTa: A Robustly Optimized BERT Pretraining Approach · Related Work: Text De-Identification and Anonymization, Privacy Models, and Adversarial Stylometry · 0.86
Cites as background 1 paper
-
State-of-the-art approaches involve using Support Vector Machines [ref] , and more recently, fine-tuned LLMs like BertAA [ref] .
On the State of the Art in Authorship Attribution and Authorship Verification · Related Work: Text De-Identification and Anonymization, Privacy Models, and Adversarial Stylometry · 0.97
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.