Study reveals AI can accurately steal password by ‘Listening’ to keystrokes

artificial intelligence,ai, tech news — Image Source : FILE Study reveals AI can accurately steal password by ‘Listening’ to keystrokes

A recent study has brought to light the potential threat of artificial intelligence (AI) tools being used to steal user passwords by "listening" to the sound of keystrokes. Researchers from Cornell University in the US conducted a study which demonstrated that an AI program, when activated on a nearby smartphone, could reproduce a typed password with an astonishing accuracy rate of 95 percent. The experiment involved training an AI model to recognize the sounds of keystrokes on a 2021 MacBook Pro laptop. During a Zoom video conference, the AI tool showcased impressive accuracy while "listening" to typing through the laptop's microphone, achieving a keystroke reproduction accuracy of 93 per cent.

The study highlights the concept of an "acoustic side-channel attack," wherein bad actors exploit the sounds generated by keystrokes to gain unauthorized access to user accounts. Researchers emphasized that many users are unaware of this threat, as they tend to underestimate the risk associated with keyboard acoustic emanations. While users may take measures to hide their screens when entering passwords, they often neglect to obfuscate the sound of their keystrokes.

ALSO READ: YouTube implements stricter guidelines for health misinformation

In order to gauge accuracy, the researchers conducted tests by pressing 36 keys on the laptop multiple times, varying the pressure and finger used for each press. The AI program's accuracy relied on identifying unique elements of each key press, such as sound wavelengths. Positioned approximately 17 centimetres away from the keyboard, an iPhone 13 mini smartphone was used in the experiment.

ALSO READ: WhatsApp introduces Forwarding Message feature for Channels on Android and iOS

The study underscores the need for greater awareness about the potential vulnerabilities associated with keyboard sounds and the importance of adopting security practices that address acoustic side-channel attacks. As technology advances, such findings emphasize the evolving challenges in maintaining cybersecurity and the need for innovative protective measures.