Hackers now steal your password with the sound of your keyboard
Type quietly or the intruder will hear you: a new AI tool allows it to identify what has been typed based on the sounds of keyboard clicks. Should you protect yourself now?
Keyboard sounds can reveal your password.
The new AI tool can tell what has been typed from keyboard sounds.
According to a study by several British researchers, the tool has an average of 95 percent accuracy so far.
The sound is picked up by the smartphones or headphone microphones around and fed into the AI.
Public places such as coffee shops or train journeys can also be a risk.
Biometric logins such as fingerprints and facial recognition help protect you.
The research team writes that this method may become popular among hackers in the future.
The new AI tool detects what you’re typing on your laptop.
AI analyzes keyboard noise.
According to a British study, the tool has a 95 percent success rate.
Credit card data, passwords, and more can all be found this way.
Preventive measures such as biometric login methods can actually reduce the risk.
We’ve known this trick since the invention of the webcam: the purpose of tape or a piece of paper is to protect the camera attached to your computer from prying eyes. After all, you never know. Will the concerned users also have to buy a quiet keyboard in the future? A new AI hack can use the sounds of keyboard clicks to learn what you’re typing.
How does this work?
Many researchers from UK universities In computer science, a new study tested whether artificial intelligence could evaluate audio signals from keystrokes. They find that the success rate of AI is 95%. MacBook Pro keystrokes recorded by a nearby smartphone were tested.
The process of a real hack mode is similar to a research test: the cyber attack is carried out using a second device, which can be a smartphone on the table, a connected headset or a microphone. This records what was typed during the session and then feeds it to an AI that is constantly learning. The bot can then use the different sounds that are produced depending on the key and typing speed of the typed text.
Deep learning is a technology used by artificial intelligence. A computer or program can learn to understand things by looking at many examples. If they are shown lots of pictures of dogs and cats, they can learn to tell the difference on their own. To do this, they process several layers of information and recognize similar patterns in our brains. Deep learning models can then use this data to automatically recognize and tag images of dogs or cats. In this article’s example, they are the phonetic patterns of keyboard strokes.
What can happen to me?
If you spend most of your computer time surfing the web and watching YouTube, chances are your risk of sharing sensitive information is low. But it happens relatively often that private messages are written in emails, social media, credit card numbers, and especially passwords. This allows the AI program to steal it.
This is how the everyday life of a hacker goes – using the example of a young man (19) who hacks into Swiss companies for money.
20 minutes / Simona Ritter
The study authors also stress that cyberattacks of this type are likely to become more likely in the future. Microphones and AI models are more common today than ever, especially in the laptop environment. These devices are portable and people use them in a variety of public places, be it in coffee shops, at work, in the library, on the train, or on school campuses. There is easy recording of keystrokes.
what can i do?
Since this type of cyberattack is new, there are no official preventive measures yet. In addition, no such cases have been reported in Switzerland yet. However, we have summarized how you can reduce the risk:
Biometric login data: Fingerprint and face recognition do not require a keyboard.
Strong passwords: Use uppercase and lowercase letters, special characters, and numbers.
Two-factor authenticationComplement your password logins with additional authentications such as SMS or biometrics.
These login methods are now standard on newer devices and are more secure than simple PIN codes or simple passwords. However, according to the researchers, one should be aware that AI is learning more and more and that the previous error rate is likely to decrease gradually in the future.
With the daily update, you stay up to date on your favorite topics and never miss any more news about current world events.
Receive the most important information directly in your mailbox every day.
“Subtly charming coffee scholar. General zombie junkie. Introvert. Alcohol nerd. Travel lover. Twitter specialist. Freelance student.”