Security researchers created an algorithm that turns a malicious prompt into a set of hidden instructions that could send a user’s personal information to an attacker.
from Feed: Artificial Intelligence Latest
Read more on Wired
