Vitalik Buterin, a co-founder of the Taifeng, launched an open experiment to invite the outside world to use any AI tool to identify a document written anonymously by him. The test revolves around the question of whether, after the generation of AI, anonymous writing on the Internet is safe enough.
Open Test Author Recognition
Buterin stated that he wanted to test an increasingly common recent claim that AI-driven writing style analysis might make it easier to restore anonymity. To complete the test, he produced an anonymous text of his own, which was given to Internet users to try to “find the real body”.
He described it as an attempt to sacrifice part of the personal anonymity to verify the real ability of AI to identify the author. According to the report, as of the time of sending, no public confirmation had been made that the document had been successfully located.
Writing style analysis due to AI speed
Writing style analysis is not a new technology. Over the past few decades, such methods have been used in situations such as author-attribution disputes, but often rely on a large number of manual comparisons and take longer.
This process is significantly accelerating as the generation of AI models is enhanced. Models can speculate on the author of the text by comparing the wording, sentences and expressions in a shorter period of time. Such identification can be easier to advance for those who have a large number of publicly written samples.
Buterin himself has long been a public publication, covering blog articles, proposals for the improvement of the ETA, research papers, forum reviews and social media presentations. This means that there is sufficient external material available for comparison.
Result or influence anonymous writing
If AI is eventually able to identify this anonymous document accurately, external concerns about privacy risks may increase further. In particular, in contexts where developers, researchers and public figures rely on pseudonyms for discussion, advances in author identification tools may change the cost of anonymous publication.
In turn, if the experiment fails, it may indicate that, even if AI ' s capabilities are rapidly enhanced, pseudonyms retain some security space. For encrypted communities, such testing is also directly related to the long-standing culture of anonymous collaboration.