Machine learning’s role in IoT security
The Internet of Things (IoT) continues to run as one of the most popular technology buzzwords of the year. With 2019 approaching imminently, the IoT landscape is only getting more populated— Gartner estimates that there will be over 20 billion connected ‘things’ by 2020.
The benefits of IoT come in plenty for many different industries, with networks of interconnected nodes and sensors producing a vast treasure trove of big data; informational gold dust serving to provide insights for a range of advanced use cases, such as;
- Helping cities predict criminal activity.
- Giving real-time information from health wearables to doctors.
- Optimizing productivity across businesses through predictive maintenance of equipment.
- Providing infrastructure for critical communications between autonomous vehicles.
That’s to name but a fraction of the frankly endless possibilities presented by IoT technology. But underlying this abundance of opportunity, the proliferation of IoT devices and large volumes of data produced carry a myriad of security threats and vulnerabilities.
It seems like every other day there is another scandalous news story describing an instance where the security of an IoT device has been compromised. Autonomous vehicles having their systems hijacked; IoT Botnet armies, such as the infamous 2016 Mirai Cyberattack which “paralyzed” the internet; children’s smartwatches being found with vulnerabilities that could enable hackers to locate their exact whereabouts- these are but a few examples of IoT security vulnerabilities.
With IoT devices growing exponentially in numbers, it is obvious that something needs to be done to better secure them. Very often, IoT devices are manufactured with little or no security controls in place. Though, fortunately, policies are now being put in place to ensure security by design.
Adding to the manufacturing issues is the fact that cybercriminals are becoming much more sophisticated in their tactics. Many are using emerging technologies such as AI (artificial intelligence) and ML (machine learning) to launch bigger and more complex attacks.
“If the bad guys are playing with a toolset that’s far more sophisticated than the ones we are using to defend, how are we supposed to win?” explained Sam Alderman-Miller, commercial director at Darktrace Industrial during the fourth annual IoT Security Foundation conference in London last week.
Evolving cybersecurity with machine learning
The huge number of IoT devices make it almost impossible for current security solutions and understaffed security teams to manually identify and stop risky activity. As such, cyber-defense programmes have started implementing AI technology in order to detect threats and vulnerabilities
These newly evolved AI-powered cybersecurity programs typically work by passively monitoring network traffic across OT (operational technology) and IT, modeling the ‘pattern of life’ for every user, device, and controller in the system.
YOU MIGHT LIKE
IoT technology gets new international standards
By doing this, it learns behaviors defined as ‘normal’ and can thus identify abnormal behaviors and potential problems at an early stage before they could cause a crisis.
AI can work as a type of ‘digital antibody’, making intelligent responses upon a threatening incident arising. The system learns from constantly evolving conditions in order to become more effective over time.
Businesses can use the technology to stay one step ahead of attacks while automating defenses, and many industries are welcoming the benefits of this.
For example, Las Vegas, a pioneer in the Smart City space, uses ML to help secure their growing network. As they deploy new technologies and projects such as autonomous vehicles, smart trash bins, and IoT sensors which control their famous City lighting, an evolved cybersecurity system using ML is vital.
Maritime is another example of an industry leveraging ML to enhance security. The digital era has created new opportunities revolutionizing the way the industry operates- from ship traffic control to automated inventory management.
But this growing reliance on innovative tools has led to new security vulnerabilities which could have devastating impacts on an industry which carries 95 percent of our trade. ML seeks to address these concerns by protecting shore-based port infrastructure and shipping fleets.
With the New Year fast approaching, it is becoming clear that ML will become paramount in the fight against rising cybercrime in the IoT era. Only AI and ML will be able to make sense of the huge amount of data required to correlate events and detect abnormal behaviors.