VK is launching an internal system on its services to protect against DDoS attacks and malicious bots, built on the analysis of user activity. The company hopes that trained AI models will use this data to distinguish between the actions of legitimate users and malicious traffic. In the future, the solution may be brought to market. However, experts doubt the possibility of adapting the system to other customers, even though big data analysis has been used to protect against cyberattacks for several years.
The company explained that AI models should distinguish between malicious traffic and real spikes in activity on the platform. By the end of the year, the system will be implemented in most of the company's projects: VK Video, OK, Zen, Mail, and others. Previously, they were protected by a set of solutions, including products from information security vendors. The company noted that AI models are trained on anonymized and aggregated data about the behavior of users of each VK service. Neural networks allow detecting malicious traffic at an early stage, even if it does not differ from organic traffic. The system should also identify parser programs and brute-forcer programs that automatically crack passwords from accounts, as well as other malicious activity on VK services.
The amount of investment in the development has not been disclosed, but a source close to the corporate software vendor VK Tech says that the possibility of selling WARP to external customers as a cloud solution is being considered.
Read materials on the topic:
VK Video Improves Search Results with Artificial Intelligence
Largest relaunch in 17 years: VK Music has been updated
VK is testing the authorization function in its services via Touch ID and Face ID
