A team of researchers from British universities has trained a deep learning model that can steal data from keyboard keystrokes recorded using a microphone with an accuracy of 95%.
Not as much as you might think. A lot of audio communication softwares these days have some sort of background noise filter, not saying thay are perfect but they have been increasingly in effectiveness and adoption. My graphics card (nvidia) even supports it and works well. That plus my quiet switches, I’m not too worried about myself, or tech savvy streamers.
The biggest target is probably the elderly and less tech savvy, who are also more likely to fall for scams, and probably have less password entropy (which would make this software more accurate).
Streamers especially are going to have to be incredibly careful going forward
Not as much as you might think. A lot of audio communication softwares these days have some sort of background noise filter, not saying thay are perfect but they have been increasingly in effectiveness and adoption. My graphics card (nvidia) even supports it and works well. That plus my quiet switches, I’m not too worried about myself, or tech savvy streamers.
The biggest target is probably the elderly and less tech savvy, who are also more likely to fall for scams, and probably have less password entropy (which would make this software more accurate).