New analysis involving scientists from Anthropic and ETH Zurich means that fashionable synthetic intelligence methods might establish the real-world identities behind supposedly nameless web accounts. The examine, revealed as a preprint on arXiv, reveals that giant language fashions (LLMs) could also be able to analyzing on-line exercise and linking pseudonymous profiles to actual people at scale.
Support authors and subscribe to content
This is premium stuff. Subscribe to read the entire article.
Login if you have purchased
