Evolving IoT security risks, impact on business continuity

ARTIFICIAL intelligence (AI) is beginning to merge with Internet of Things (IoT) devices, creating what is known as AIoT. The integration of AI into IoT assets enables them to collect, analyze and act on data autonomously.
In a typical IoT setup, connected devices (like sensors) gather and send data to be processed, often by centralized systems. However, by embedding AI into IoT, these devices can:
– Make decisions locally. AI algorithms enable devices to process data in real time without relying on cloud servers, allowing faster responses.
– Predict and optimize operations. AIoT systems can predict trends, such as machinery failure in Industrial IoT (IIoT) and automatically schedule maintenance before issues arise, leading to predictive maintenance.
– Enhance automation. AIoT devices can automate processes based on intelligent decision-making.
It is important to keep in mind that AI algorithms are not immune to manipulation. Cybercriminals can exploit vulnerabilities in AI models, poisoning the data to force IoT devices into unsafe behaviors or decision-making. For example, compromised AI models used in IIoT environments could cause sensors to give false readings, disrupt operations or damage equipment.
In addition, AIoT devices are inherently complex. Rule of thumb in the cybersecurity realm: the more complex the device, the harder it is to secure. The combination of IoT and AI means that security must be applied to the hardware, firmware, software, communication protocols and AI models, as each of those can be targeted separately by cybercriminals. The best practice would be to implement AI security measures in a very early stage. As AI becomes integrated into IoT devices, organizations and especially CISOs should focus on AI model integrity and explainability, ensuring they are resilient to tampering.

Growing threat to IoT networks