Artificial intelligence and machine learning are making waves in the maritime industry, enabling shipping companies to predict and react to emerging cyber threats quicker and more effectively than ever before. What are the possibilities and challenges?

Over the past decade, ships have gone from being vessels filled with analogue equipment and machinery to highly automated floating computers, increasingly dependent on software-based control systems.

The digital ship – connected to the Internet – is transforming shipping operations, allowing supply chains to be more efficient and transparent.

However, in a shipping industry facing the stark new reality of complex and multifaceted cyber threats, ship connectivity also brings vulnerabilities.

From 2017 to 2020, cyberattacks on the maritime industry’s operational technology (OT) systems increased by a whopping 900%. Needless to say, such a massive volume of attacks on increasingly complex and interconnected systems makes it impossible for humans to connect the dots on their own.

Even advanced cybersecurity systems need to work extra hard to detect and block the tsunamis of malware and similar unwanted traffic that modern, IoT-connected ships are up against today.

Fortunately, machine learning is levelling the cybercrime playing field.

How can machine learning help shipping beat cyber attackers?

First of all, what is machine learning? A subset of artificial intelligence (AI), machine learning is a set of techniques and technologies using algorithms to examine large volumes of information or training data to discover unique patterns, which it can then analyse, group and use to make predictions. Machine learning can learn from data and make decisions without the aid of human interaction.

With their ability to sort through millions of datasets, recognise patterns and identify anomalies, machine learning systems are increasingly being used to proactively uncover data breaches or intrusions.

Cybersecurity solutions powered by machine learning use data from past cyberattacks and threats, learn from the data, and identify and respond to similar threats.

An adaptive monitoring solution that leverages machine learning eliminates any threat in its infancy and sends off alerts to the right people before it can compromise your systems and operations.

Challenges

Obviously, this AI-powered approach to maritime cybersecurity saves a lot of resources and helps detect new threats quicker, allowing necessary updates and patches to be released more proactively.

However, does machine learning represent a silver bullet for maritime cybersecurity? No – simply because there is no turnkey solution to building a cyber-resilient environment onboard your ships.

You can’t simply ‘turn on’ machine learning and expect it to add a bulletproof layer of defence to your ships’ infrastructure. For AI algorithms to be able to detect anomalies and threats, they need to be fed months of event logs to set a baseline of normal performance. From there, the machine learning system can identify sophisticated patterns in the logs, and recognise malware or hacker activity that would otherwise take weeks or months to detect – if detected at all. 

Another challenge is the fact that cybercriminals themselves will use AI, to learn and adapt to your cyber defence measures. Moreover, as cybercriminals inevitably get their hands on AI, so you as a shipping manager must use it to safeguard your IT environment. 

Maritime cybersecurity service powered by machine learning

Depending on yesterday’s countermeasures in this ongoing cyberwar at sea is like bringing a knife to a gunfight. Machine learning enhances cybersecurity in multiple ways, allowing you to detect ever more complex threats from ever more sophisticated attackers. 

To stay ahead and stay safe, you should invest in an enterprise-grade cybersecurity service that leverages machine learning to ensure that your fleet IT infrastructure remains rock-solid.

Editor's note: This article was originally published in June 2019 and has been revised and updated for accuracy and comprehensiveness.