Scary AI recognizes passwords by the sound of your typing

Dec 9, 2023 ·

AI technology is being used to detect passwords by analyzing the sound of typing. This new system developed by researchers from the National University of Singapore (NUS) and Pennsylvania State University (PSU) uses machine learning models to identify passwords with up to 85 percent accuracy.

The technique works by using microphones to pick up on subtle differences in how users type their passwords, such as the speed, rhythm, and pressure applied to the keys. It then processes the audio signals through a machine learning algorithm to determine the password.

The researchers tested their system with 37 different keyboard patterns, including QWERTY, Dvorak, Colemak, and others. They found that it was able to distinguish passwords from each other with up to 85 percent accuracy.

The research team believes that this technology could be used to improve security and authentication systems. For example, it could be implemented into existing password-protection mechanisms to automatically detect when someone is trying to guess a password or break into an account. It could also be integrated into smart home devices, as well as cars and other connected objects, to strengthen security.

The research team plans to further refine the system, and hopes to eventually develop a system that works with any type of keyboard layout. With the right tweaks, the system could become a powerful tool for reducing fraud and improving security measures.