publive-image

AI algorithms are able to find patterns across large datasets and to detect anomalies speedily. In other words, the advantage of an AI algorithm helps to create a baseline model of what's normal in an environment and then flagging and inspecting anomalies to that baseline, which may be IoT threats. This can be seen in different environments, from ones as small as an individual residence up to the largest businesses.

For example, there are many IoT devices in your home. The light bulbs can only talk to the light bulb controller, and the controller will then talk to the app which is used to control the lights. Similarly, the speakers can only talk to the central controller unit for the speakers, which can then talk to my apps. The doorbell should only speak to its related app. This is a standard baseline. There is also an innovative thing that when you put up an app to flash the lights with the ringing of the doorbell, and you can also set an appropriate light to come with playing the music.

The above example of IoT threat detection ranges to both user and group of machines. According to observable data, an AI algorithm able to find out which operators need to be able to configure IoT devices, what times they usually do configuration, where that configuration request begins and so on. Likewise, an AI algorithm is able to find out usual communication paths from IoT devices to controllers, and also between IoT devices. Therefore, threats are the irregular activities outside of what the AI system has seen before.

Assume that it is a definite degree of human interaction with the AI to create judgment calls about whether an irregular activity is correct. The AI at home doesn't have the authority to tell that the lights shouldn't talk to the speakers. But it needs approval from the operator. This is a good thing as it helps control against irregularities or an IoT threat.

The same type of authority should be implemented in an enterprise setting to provide proper IoT threat detection. You have to appoint the right person or people to make judgment calls about probable unusual behaviors and set up new rules in place for the AI algorithm and on it, the security of the IoT system will depend.