Vitalik Buterin, a co-founder of the Taifab, launched an open test to invite Internet users to find an Etherak document, written anonymously by him during the decade and publicly released. The test directly turned the discussion around AI text analysis and network anonymity into an observable experiment.
Buterin suggests that there has been a saying that AI writing analysis makes online anonymity increasingly difficult to maintain. On the basis of this judgement, he decided “to experiment with part of his anonymity”. However, he did not publish the name of the document, nor did he give more direct location.
The document range is not yet public
According to Buterin, this is an Etherwood file of 'medium importance'. He estimated that there were approximately 200 to 2,000 documents of equal or higher importance in the ETA. This means that the outside world is confronted with a wide-ranging search task, rather than a simple comparison between several candidate texts.
As of the time of sending, there is no publicly available information indicating that the target document has been successfully confirmed. As a result, the test remains open and the results cannot be judged for the time being. Unless Buterin himself or other reliable sources confirm the answer, it is difficult for the outside world to verify whether a given speculation is accurate.
Focus on textual recognition
At the core of the test, it points to the recognition of the body, i.e. the inference of the author ' s identity by writing style, word habits and sentence structure. The methodology is not new, and researchers and investigative agencies have been using it for many years. The difference is that the new AI tool allows for faster processing of larger aggregates of text, with a significantly more efficient identification than manual matching.
If AI is able to target anonymous authors on a larger scale, pseudonym writing space on the network may be further compressed. This issue is particularly sensitive for encrypted communities, as many researchers, developers and governance participants have long relied on semi-animous or pseudonyms for public discussions.
Instead of his recent privacy calls,
Buterin's recent concerns about AI security and privacy have also continued. He had previously argued for greater use of local priority programmes for the AI tool, on the grounds that cloud deployments were more likely to lead to a risk of data exposure, information leakage and manipulation.
At the Ethera level, he has also spoken in recent times about the direction of privacy improvement. Previous disclosures envisaged the reduction of the risk of metadata leakage and review through the abstraction of accounts, keyed nonces and access-level optimization. At the same time, he mentioned that AI could be used for formalization certification, helping to develop codes to enhance their authentication.
At present, this open test is pending. However, whatever the outcome, it is possible to provide a new sample of a larger problem: the extent to which the open community retains anonymity when AI is better at identifying authors from the text.