AI increases the risk of hacking smart devices: Safety tips

From asking your smart speaker for the weather to receiving personalized advice from your smartwatch, devices powered by artificial intelligence (AI) are increasingly streamlining our daily tasks and decisions. This technology has entered our lives in subtle ways.

Manufacturers collect vast amounts of user data to ensure the responsiveness and personalization of these smart devices. However, this can put users at risk of exploitation by malicious agents such as hackers looking to steal data.

As AI becomes more pervasive, consumers will also need to become more knowledgeable. If you want to enjoy the benefits of smart everyday devices, you need to be aware of the necessary safeguards to protect yourself from cyber-attacks.

A smarter internet of things

The Internet of Things (IoT) was born when we started connecting physical everyday devices like refrigerators, vacuum cleaners, and doorbell cameras to the internet. It is estimated that there are approximately 17 billion IoT devices in the world today.

IoT devices that pre-existed AI typically have simpler and more static functionality, which poses lower data privacy and security risks. These devices can connect to the internet and perform specific tasks programmed to remotely turn off the lights or set the thermostat.

However, it could not learn from user interactions or adapt its functionality over time. Manufacturers integrate his AI into her IoT devices so that they can “understand” and better respond to users' needs and behaviors.

For example, smart speakers may collect behavioral information by listening to conversations around them. This allows us to better understand your preferences and commands, tailor our responses, and provide you with more relevant content and suggestions. Ultimately, this will improve your experience and make your device more useful.

However, it is also less secure. As AI becomes embedded in such devices, it actually opens up a new set of avenues (known as the “attack surface”) for cybercriminals. For example, a hacker could use inputs that intentionally cause her AI within the device to malfunction. You can also “contaminate” the training data of an AI model to make it behave in a particular way.

Additionally, malicious attackers can obtain AI training data through model inversion attacks. If an AI model is trained on private or sensitive data, replicating this model can potentially expose information that should be kept confidential.

Manufacturers should do more

IoT devices have long been vulnerable to hackers due to lack of passwords, lack of encryption, or outdated software. With this in mind, smart device manufacturers that prioritize security implement strong encryption, provide regular software updates, and ensure secure data management and transmission.

However, users are often unaware of how vulnerable their devices are, what kind of data they collect, and where it goes.

There is an urgent need for industry standards to ensure that all devices meet minimum security thresholds before being brought to market.

Manufacturers must provide detailed guidelines on how the data collected will be processed, stored and protected. You should also explain the measures you take to prevent unauthorized access and data breaches.

Governments and industry are recognizing the risks and invisible threats posed by AI. We have already seen that this can have significant negative effects if misused. That is why AI regulation laws are being drafted and implemented in Australia and around the world.

In the meantime, consumers must remain vigilant and take proactive steps to ensure their digital lives bring more good than harm.

How can I protect my device from cyber attacks?

First, check all the devices in your home that are connected to the internet. Be sure to identify AI-powered capabilities, such as learning user behavior or processing large datasets. These are common in smart speakers, home security systems, and advanced wearable technology.

Next, examine your device's capabilities and disable irrelevant or unnecessary AI features. These simple steps can prevent AI from collecting your personal information and potentially leaking it.

Third, when purchasing a device, check the manufacturer's security disclosure information. They are often listed on websites under titles such as “Privacy,” “Security,” or “Product Support.” It can also be found in the user manual and, in some cases, directly on the product packaging.

Be sure to understand what types of AI technology your device uses and how your data is collected, processed, stored, and protected. What are the safeguards? Did the manufacturer use industry standards or adhere to strong security guidelines such as the European Union's data protection regulation GDPR?

Security disclosures can vary widely in terms of clarity. The technical details can be difficult to understand, but the information in the Australian Government's Consumer Data Rights Guidelines can help guide your decision.

Asking these questions will help you choose a device. Rather than relying solely on price, it may be best to choose a manufacturer with a proven track record for security.

Finally, keep your IoT devices updated. If your device asks you to install an update, do it immediately. This ensures that security loopholes identified by the manufacturer are properly implemented and opportunities for cyber attacks are blocked.

These good habits will go a long way in ensuring your privacy is protected.

Chao Chen receives funding from a variety of organizations, including industry partners for industry-led research projects and government funding from ARC and CSIRO.

Kok-Leong Ong receives funding from a range of organizations including industry partners for industry-led research projects and government funding from NHMRC, MRFF and CSIRO.

Lin Li does not work for, consult, own stock in, or receive funding from any company or organization that might benefit from this article, and has no relevant affiliations other than an academic appointment. has not been made clear either.

/Courtesy of The Conversation. This material from the original organization/author may be of a contemporary nature and has been edited for clarity, style, and length. Mirage.News does not take any institutional position or stance, and all views, positions, and conclusions expressed herein are solely those of the authors.