A recent study by Anthropic and ETH Zurich suggests modern artificial intelligence tools are capable of connecting anonymous digital accounts to actual people. The research, posted on arXiv, found that large language models can sort through online texts, identifying writing traits and interests, then match them to known individuals on a large scale-an ability that used to demand significant manual effort.
In their experiments, researchers attempted to link users across different platforms, such as connecting anonymous Hacker News participants to LinkedIn profiles and matching unrelated Reddit accounts to the same person. Using just publicly accessible information, the AI system significantly outperformed older manual techniques, accurately linking accounts with high precision and recall, while traditional approaches were much less successful.
This advancement means that tasks previously handled by experienced investigators can now be rapidly performed by AI, which could threaten online anonymity for users like reporters or activists. Even though this work relied on controlled data and has not yet been peer-reviewed, it has already increased concerns among privacy advocates and technology specialists. The researchers stress a growing need to rethink personal data sharing online and to create stronger privacy safeguards as artificial intelligence continues to improve.
The ainewsarticles.com article you just read is a brief synopsis; the original article can be found here: Read the Full Article…
