In the last 12 months, 60% of Australian organisations experienced a ransomware attack. This is according to Telstra’s Cyber Security Report 2017, which also found that ransomware was the number one type of malware downloaded in the Asia Pacific region during 2017.
With ransomware being one of the leading cyber security threats to Australians, you don’t need to have been a victim to understand that cybersecurity is one of the most pressing challenges businesses across the nation are facing. Analysing yesterday’s security incident no longer enables you to predict and prevent tomorrow’s attack.
A significant challenge for businesses is that legacy antivirus technology is too slow to stop cyber-attacks in time. Attacks are often weeks or months old by the time they are discovered, as many security point solutions only store a few days or weeks’ worth of data. Throughout the last year, 24 percent of Australian organisations experienced a ransomware incident on at least a monthly basis and it took five hours or more to recover. In addition to enterprise ransomware attacks, we are also witnessing criminal groups leveraging sophisticated exploit tools and living-off-the-land techniques to increase dwell time. In the cybersecurity industry today, one of the particularly challenging areas to address is combating malware-free intrusions.
It is clear that businesses need new ways to defend themselves again continuous and evolving threats. Conversely, many of the traditional approaches to security are now ineffective. In today’s modern IT landscape, Australian organisations need to look to more next-gen solutions to combat effectively against modern threats. Replacing traditional, legacy antivirus (AV) technology with a more sophisticated approach that combines Artificial Intelligence (AI) and behavioural analytics is a key component of this.
Putting Artificial Intelligence into Context
AI has become a prominent industry buzzword but confusion remains as to how it is being applied effectively in the context of information security. The 2018 Gartner CIO Agenda Survey found that CIOs are increasingly adopting AI in their organisations, predominantly to boost the customer experience or fight fraud.
For most security scenarios, AI enables capabilities that go far beyond identifying known threats. AI models can determine a file’s maliciousness with no previous knowledge of the file, relying instead upon analysis of the file’s innate properties. With sufficient quality data available, AI techniques easily outperform traditional signature-based or indicator of compromise (IoC) based prevention approaches, which retroactively seek out the artifacts an attacker leaves during a breach.
However, there are many breach scenarios that happen today that require behavioural analysis to isolate threats based on observing the actions taken. According to an industry study, over 60% of intrusions don’t actually involve any malware, but instead leverage stolen credentials and ‘living-off-the-land’ techniques like use of powershell and legitimate Windows tools.
Adding behavioural analytics to a prevention strategy built on AI/machine learning is the key to raising the bar in securing corporate networks.
Finding the Right Solution
In cybersecurity, speed and scale matter. This is exactly where AI adds a significant advantage to enterprise security methods. However, not all AI tools are created equal. For many CIOs who are new to integrating AI into their strategy, here are a few questions to help them evaluate solutions:
- What is the volume of data sets the algorithms were trained against? This is critical for effective detections and reducing false positives.
- Is the data collected from a global, cross-section of industries? Different industries experience different types of security events so having a cross section of verticals will help build more holistic views of cyber threats.
- How do you retrain algorithms and classifiers? The continuous, real-time updating process is crucial for staying ahead of fast-evolving threats. This is one area where cloud solutions have a distinct advantage because their enable frictionless, instant updates.
As we reflect on the way organisations around the world have been impacted by breaches this year, it’s clear that traditional approaches to security have failed. Look no further than WannaCry, which saw more than 300,000 computers across more than 150 countries get locked up by the ransomware. Shortly after this, WannaCry’s evil twin brother, NotPetya, had a large impact around the world, bringing down hospitals and healthcare organisations, manufacturers and logistics companies, as well as corporate firms.
The impact of these attacks point to the fact that traditional security measures deployed in each organisations’ architecture simply did not stand up. Those affected are starting to report material impact on their financial performance and US congress members are inquiring about the effects on drug supplies following outages attributed to NotPetya. The real impact is now becoming clearer in the months post incident. With this in mind, now is the time for business leaders to be thinking seriously about the role of AI in enterprise security strategy.