IBM develops malware 'Deep Locker' to bypass security software and attack only specific targets
In order to study software to counter new viruses and malware damage by IBM, we developed a malware " Deep Locker " that allows you to attack only specific targets while avoiding security countermeasure software.
Deep Locker: How AI Can Power a Stealthy New Breed of Malware
Researchers Developed Artificial Intelligence - Powered Stealthy Malware
In recent years, the development of AI technology has been remarkable, and the development of security software that protects computers from various threats using AI is under way. However, backing back, AI technology may be used for viruses and malware, and by breaking through various security measures easily, there is a possibility of enormous security damage.
Therefore, in order to demonstrate the threat of malware using AI, the research team of IBM Basic Research Laboratories developed a malware "Deep Locker" that prevents anyone from being able to detect existence until it hurts the target.
Normally, viruses and malware attack unspecified number of computers. In the case of Deep Locker, we can narrow down the target to specific people and organizations using speech recognition, face recognition, etc. Furthermore, it can behave as ordinary application software such as video conferencing software until attack starts, and has the function of protecting you so that it will not be detected by security countermeasure software.
IBM's research team is also conducting demonstration experiments using DeepLocker to check whether it can actually damage only the target person's PC. In experiments, it seems that Dean Locker incorporated Wanna Cry of Ransomware and sent it on the network. Then, Deep Locker invaded a lot of PCs and confirmed owners using PC's camera. After that, only Deep Locker that has infiltrated the target PC has executed Wannacry, and it is proven that malware works properly. The picture used by Deep Locker to recognize the PC owner 's face is that it uses only the photos posted on SNS. As a result, it has also been shown that it is possible to create malware that can only be affected by the relevant person without having to get close to the target person in advance.
The research team says "We will advance research on new defense measures assuming attacks by malware using AI", and we are going to make use of Deep Locker as an experimental material for new security countermeasure software .