Caution! AI Tool Can Capture Passwords through Keystroke Sounds

- Technology - August 17, 2023
AI Tool
AI Tool

Caution! AI Tool Can Capture Passwords through Keystroke Sounds


An emerging AI technology is raising cybersecurity concerns as it can potentially extract passwords from the auditory signals of keystrokes made during virtual meetings like Zoom, or any application with microphone access. The tool demonstrates an impressive accuracy rate of around 90% in deciphering these sensitive inputs.

Recent research highlights a worrisome vulnerability: your passwords could be compromised while typing during a Zoom call. Cyber attackers have harnessed artificial intelligence capabilities to ‘listen’ to keystroke sounds, marking a significant advancement in sound-based cyber threats.

The Study’s Findings

Conducted by researchers from Durham, Surrey, and Royal Holloway universities, the study unveiled that AI algorithms can decode distinct typing patterns based on auditory cues. The study, published on August 3rd, underscores the growing risk of sound-triggered cyberattacks due to the rising usage of video conferencing platforms like Zoom and the widespread presence of microphone-equipped devices.

AI’s Remarkable Abilities

The AI tool in focus exhibited a remarkable ability to interpret text, including passwords, from the auditory feedback of keystrokes. This tool, which underwent thorough analysis by the research team, achieved an accuracy level surpassing 90%, indicating its potential threat to data security.

The Research Process

The researchers embarked on a rigorous training process for the AI model. By pressing each of the 36 keys on a MacBook Pro, they recorded the corresponding auditory keystroke sounds. These sound samples were then integrated into the AI model, enabling it to accurately recognize the distinct sonic signature associated with each key.

Impressive Results

The AI model’s performance was remarkable. It showcased an accuracy rate of 93% in deciphering keystrokes from recorded MacBook typing sounds, particularly through the Zoom video conferencing software. Furthermore, this accuracy rate surged to 95% when the keystrokes were captured using an iPhone 13 mini.

Protecting Against Sound-Based Cyber Threats

As the adoption of virtual communication platforms surges, safeguarding sensitive information becomes paramount. Implementing protective measures such as muting microphones when not speaking, using background noise suppression tools, and enhancing data encryption are crucial steps in mitigating the risk of sound-based cyberattacks.


The emerging AI tool’s ability to extract passwords through the analysis of keystroke sounds is a concerning development in the realm of cybersecurity. The research underscores the importance of staying vigilant during virtual meetings, employing cybersecurity best practices, and adopting technological countermeasures to safeguard personal and sensitive information. As the digital landscape evolves, prioritizing data security remains a collective responsibility.