Malboard: Now Hackers Can Make Use Of Our Keystrokes To Hack

Recently, researchers have developed a new type of cyber attack that is capable of mimicking the identity of the user via their keystrokes. Unfortunately, the continuous evolution and the high level of sophistication of cyber attacks recently lead to the reality where antivirus products base on signature are not enough anymore.

Behavioral patterns based verification system

We have also look into using behavioral patterns (like mouse movements or keystrokes) to verify our identity. However, researchers from Ben-Gurion University of the Negev (BGU) said, this is not foolproof.

This Wednesday, the BGU's team announced that they have successfully developed a new type of cyber attack - Malboard. This new attack could evade all of those detection products on the market that use keystroke characteristics to verify the identity of the user.

How they did it?

In an online publication, the BGU's team pointed out how the hacker could use a fooled keyboard to send, generate malicious keystrokes that mimic the user.

The researchers used keyboards from Dell, Lenovo, and Microsoft for their research. The goal was to trick DuckHunt, TypingDNA, and KeyTrac ( authentication systems base on behavioral patterns). These software use machine learning and algorithms base on AI to verify our accounts with our keystrokes. Still, people could use those same algorithms to fool them.

The team at BGU gathered the data to develop Malboard from the result from 3 keystroke tests of 30 participants. Then, they fed this data to the underlying AI algorithms and database of the attack, which will mimic the keystrokes of the participants to go against the authentication systems. From 83% to 100% of the tests, DuckHunt, TypingDNA, and KeyTrac were fooled.

Doctor Nir Nissim of BGU said that Malboard would be useful in 2 scenarios: remote attacks by hackers remotely, or inside attackers who are capable of launching Malboard on the internal system's keyboard.

The paper of the team also suggests using detection modules to improve the verification system base on the keyboard, including keystroke sounds, detection of typographical errors, and monitoring power consumption.