New acoustic attack determines keystrokes from typing patterns

Researchers published a proof-of-concept acoustic side-channel technique that uses keystroke sound signatures and timing intervals to predict typed text, achieving an average 43% accuracy even under noisy conditions and across different keyboards; the attack requires target-specific typing samples (which could be captured via malware, browser extensions, compromised devices, or concealed microphones) and is limited by factors such as silent keyboards and variability in individual typing patterns.