AI is Making Smart Devices Easier to Hack – What Can Be Done?

IoT. Image by pixabay

This post is also available in: עברית (Hebrew)

Our everyday routines are becoming increasingly flooded with AI-powered smart and ultra-connected devices (Internet of Things), and manufacturers are gathering incredible amounts of user data to ensure these smart devices are responsive and personalized. However, this puts users at risk of exploitation by such as hackers looking to steal their data.

While IoT devices existed before AI, they had simpler functionalities and posed lower data privacy and security risks – they could connect to the internet and perform specific tasks they were programmed to do but couldn’t learn or adapt their functionalities over time. Manufacturers then began integrating AI into IoT devices to help them better cater to user needs and behaviors, which enhances the user experience and makes the device more useful. However, it also makes it less secure.

The fact AI is now embedded into many smart devices opens a new collection of pathways (called an “attack surface”) for cybercriminals, who in turn can use inputs that deliberately cause the AI in the device to malfunction or make AI models misbehave in specific ways by “poisoning” their training data.

According to Techxplore, there are many reasons IoT devices are vulnerable to hackers, including the lack of passwords, lack of encryption, or outdated software. Therefore, experts are calling for smart device manufacturers to prioritize security and implement strong encryption, provide regular software updates, and ensure secure data management and transport.

There need to be industry standards that ensure all devices meet a minimum-security threshold before they reach the market, with manufacturers providing detailed guidelines on how the collected data is processed, stored, and protected.

Meanwhile, users must take proactive measures to ensure their smart devices are not causing more harm than good. Experts recommend users first review the devices in their homes that connect to the internet and identify AI-powered features (common in smart speakers, home security systems and advanced wearable technology) and disable irrelevant or unnecessary AI features.

Finally, experts recommend checking the manufacturer’s security credentials when buying smart devices, and keeping those devices up to date when they recommend installing an update (which ensures treatment of any vulnerabilities discovered since the last update, and to prevents cyberattacks).