Back to Insights

How AI Can Cover The Security Skill Gap?

Managed services 10/09/2020 - 15:08 by Swami Nathan

Although AI is the newest culprit, concerns about human replacement technology date back to the 2nd Industrial Revolution when the economy shifted, and farmers switched to more manufacturing and railroad jobs. Society was worried that they would see the end of the days when real people produced results, not machines.

The Digital Transformation is in full swing, but does cybersecurity continue to keep up? Last month alone, at least 99 cybersecurity incidents were recorded. Making it the third-largest monthly total by several security incidents in the year – indicating significant room for growth in cybersecurity systems. From banks to social media platforms it is obvious that breaches of security are a universal threat to all IT infrastructure.

While organizations scramble for talent, experts look to another type of intelligence to fill the ranks of cybersecurity.

Using AI for Research

AI and machine learning technologies remove complexities for the security team from the experience of threat detection. Junior analysts are allowed to do investigations using these technologies, freeing up the attention of senior analysts so they can focus on solving bigger problems.

Machine learning assists in the process of investigation by focusing on specific events linked to a user or device. If the user or device in question shows red flags, AI can determine if the specific behaviour goes beyond established thresholds and describe the underlying behaviour to cybersecurity professionals.

AI and machine learning allow security analysts to gather from their logs the following information:

  1. Identify users and computers with anomalous behaviours.
  2. Determine whether an account belongs to a person or a computer program.
  3. Identify peer groups based on user behaviour and IT environment interactions.
  4. Render host-to-IP mapping automatically.

Legacy Resource Limits

Over the past few years, as the cyber-threat environment has changed, businesses have switched to cybersecurity firms providing tools for security information and event management (SIEM). However, the increase in cyber-attacks, the lack of trained security analysts and the increasing number of defensive devices have created operational problems with legacy SIEM vendors. For instance, SOC teams complain about wasted time chasing false positives, being unable to capture unknown threats, avoiding distributed attacks and having to manually investigate and solve problems. Enterprises may also face unnecessary costs of logging.

Security analysts rely most of the time on principles of association to identify potential threats. Analysts need to know what they are looking for in advance to get them to work. For example, a log rule for identity and access management (IAM) might involve raising an alert when creating and deleting the same user account within 24 hours.

Security teams equipped with AI technologies can remove the need for prior knowledge of attacker tactics and techniques as opposed to legacy resources. Using machine learning, security teams can pre-process logs to identify anomalous user and asset activities and combine them with other data sources.

Why Raw Logs is not fine?

Modern enterprises have a large and increasing number of endpoint devices, applications and services which make it difficult to manage security and IT operations with network monitoring and logging alone. Unfortunately, it can take hours for security researchers to manually sift through a wider variety of events. Moreover, dependency on raw logs goes against the top three goals listed in the above survey.

Raw logs restrict how much detail about incidents is presented to observers and contribute to false positives. However, machine learning, combined with the inclusion of contextual data sources and the knowledge of threats, will enrich log data.

Semi-autonomous safety platform

AI’s deep learning and cognitive computing components can help detect malware, intrusion, fraud and even user and computer protection risk analysis.

Deep learning (DL) AI can process and learn from unstructured or unlabelled data, setting them apart from other methods of machine learning which require structured and labelled information to be fed. DL thrives on massive volumes of data, which is an atmosphere that SOC can offer. In the same vein as DL, cognitive computing strives to function like the human brain, incorporating different Artificial Intelligence strategies in machine learning, natural language processing and human interaction to develop knowledge and make autonomous decisions on its own.

Similarly, professionals use analytics to weed out anomalies in their network, an AI-infused framework for security information and event management (SIEM) as it improves threat detection through deep learning methods. The cognitive computing capabilities of your SIEM build a cybersecurity system which is continuously learning and adapting to threats. The AI offers guidance once an intrusion has been identified, which helps analysts take action quickly. This frees engineers’ time to transfer their attention to other goals within the SOC, rather than directing the hands-on SIEM.

Not all security solutions are created equal. Organizations need to avoid anticipating AI and machine learning being integrated into security technologies and platforms. Not only will AI play a vital role in protecting the business from an attack, but it will also ensure that the security teams don’t waste their time on tedious tasks.

Over many decades, as society looks back on how technology has evolved, we hopefully find that AI has become one of our greatest strengths and allies in cybersecurity and beyond.

The only way to protect what you’ve worked hard to build is to be vigilant when it comes to cybersecurity. If you’d like to know more about how your business can benefit from managed services, just give us a call, we are here to help.

Although AI is the newest culprit, concerns about human replacement technology date back to the 2nd Industrial Revolution when the economy shifted, and farmers switched to more manufacturing and railroad jobs. The society was worried that they would see the end of the days when real people produced results, not machines.