Rodney Buang Sebopelo
North West University – MSc
From students through to tenured professors, everyone on a university campus is on the internet. Whether it is on their laptops and desktop PCs or on smartphones, all of these people – sometimes thousands at a time – are connected to each other through the university’s intranet. Increasingly, they are connected via WiFi rather than through a cable plugged into the back of their computer. And this makes everyone vulnerable. A cyber attacker could enter someone’s device and fiddle with their information – for profit or malice – or take control of their device.
As smartphone usage has increased, so has cybercrime. On a mobile ad hoc network (Manet), such as a university’s WiFi intranet, it is impossible for a person to anticipate or even identify when a device has been compromised. A Manet’s dynamic nature means that it is even more vulnerable to attack.
To protect the safety of everyone on the Manet, it is imperative to secure possible criminal loopholes and identify a compromised device as soon as it has been taken over. But there are many ways mobile devices and remote connectivity can put the operation of a system in danger – too many for a human to keep an eye on. These danger-detection strategies include steps to identify technologies and practices such as unauthorised network access, unsecured or unlicensed applications, viruses, malware and others.
With this in mind, researchers have identified a set of security controls and countermeasures to fight off mobile threats. To provide secure communication on the network – which is often where the attacks take places – North West University researchers have created security measures to check that everyone’s device is still within their control. This involves checking the behaviour and cooperation of each device, and that they allow information to travel safely from one node to another. Essentially, if a device has gone rogue, it will not comply with this spot check, or will not be able to share information safely between nodes.
But with thousands of devices, it is not possible for a human being to do the checking and verify whether devices are still within the control of their owners.
This is why researchers are turning to artificial intelligence (AI) to detect malicious people or software on a network.
While AI has become a buzzword in tech circles – it’s often touted as a solution to any data problem – it is actually a very specific area of technology. AI refers to any device that can perceive its environment and take actions that maximise its chance of achieving some goal – whether that is credit card fraud detection or protecting a virtual network. Ultimately, it is about problem solving.
In the case of Manets, researchers at North West University are using AI to identify or detect malicious activity before it affects the networks or other nodes. They do this by using machine-learning techniques, so that their software learns about malicious devices’ new behaviour as it stops them. This is why AI is one of the best way to hunt for rogue devices and malicious attackers.
AI provides a quick response to compromised devices and malicious adversary attacks. This system’s tailored algorithms classify intruders and flag cooperative mobile devices.
The machine-learning algorithms that I am using in my research gather information about the behaviour of all the mobile devices on the network, and classify them as malicious or benign. The algorithms flag which devices are dodgy and can predict which devices could go rogue. This system is also easy to test and use.
This silent protector will be able to monitor your device, and make sure that it is still in your control, without you even knowing it is there.