Researchers at Cornell University have developed an AI model capable of stealing passwords by listening to keyboard keystrokes with up to 95% accuracy. The AI model, trained on the sound, intensity, and timing of keystrokes, can be deployed on a nearby device with a microphone. The team tested the model’s accuracy during Zoom and Skype calls, achieving 93% and 91.7% accuracy respectively. The researchers suggest using features like Windows Hello and Touch ID or investing in a password manager to avoid such attacks.
Read more at Digital Trends…
