Cybersecurity and AI
For a long time, cybersecurity was a war among humans. The protection mechanism was largely a “seal the borders” approach, via firewalls, proxies, antivirus software, access controls, dynamic passwords, and so on. However, these old methods are beginning
- PDF / 167,876 Bytes
- 11 Pages / 439.37 x 666.142 pts Page_size
- 84 Downloads / 253 Views
6 Cybersecurity and AI For a long time, cybersecurity was a war among humans. The protection mechanism was largely a “seal the borders” approach, via firewalls, proxies, antivirus software, access controls, dynamic passwords, and so on. However, these old methods are beginning to seem inadequate as the battlefield has changed to human versus machine. With the increasing number of entry points into the enterprise systems landscape and more connected devices as endpoints, news about security breaches are appearing more frequently than before. Cloudification, IoT, and BYOD (bring your own device to work) are all giving rise to microenvironments that contain a lot of sensitive data about the user. If these devices fall into the wrong hands, this could certainly lead to grave consequences. The situation is further alarming when you consider the newer type of attacks, which are mostly machine engineered. For example, adaptive malware is created using machine learning techniques. This type of malware can infiltrate a system, collect and transmit data about that system, and remain undetected for days. Applying the old approach of gathering information about data breaches, malware types, and phishing activities and then creating signatures is no longer potent enough to handle the next generation of cyberattacks. The threat is real and also magnified, primarily due to increasing digitization all around us. Standard automated threat-detection systems, although improving in sophistication, find it challenging to react quickly to unanticipated or newly formed threats. © Soumendra Mohanty, Sachin Vyas 2018 S. Mohanty and S. Vyas, How to Compete in the Age of Artificial Intelligence, https://doi.org/10.1007/978-1-4842-3808-0_6
144
Chapter 6 | Cybersecurity and AI A new approach is required to continuously monitor the large number of factors and detect what constitutes abnormal activity. In short, they need to apply self-learning techniques to spot what could be a malicious activity without being told what to look for! This new approach could be similar to our body’s immune system, where the white cells and antibodies are continuously scanning and neutralizing any organism that does not fit the normal functioning patterns within the body. This is where AI comes into play. Machine learning algorithms can recognize potential security breaches or attacks by continuously observing what is an abnormal behavior, and if given the authority and the right credentials, they can automatically shut down systems under perceived threat, thereby reducing or isolating risks to the entire enterprise. There are a few challenges with this approach though. Machine learning lacks the general knowledge required to distinguish real threats. This leads to too many false positives, too many false alarms, and frequent shutting down of systems. For AI to play a decisive role in cybersecurity, it must be treated with caution and applied to problems it can solve. A hybrid human-machine collaborative approach to cybersecurity could be a potential solution.
Data Loading...
