Skip to main content

Can AI processing at the edge help maintain our privacy in smart homes, cities, and beyond?

(Image credit: Image Credit: John Williams RUS / Shutterstock)

The rapid progress in artificial intelligence, smart devices, and smart cities promises to revolutionise the way we work, live, and connect. However, recent scandals surrounding the handling of user data have prompted a wave of privacy concerns. The smarter a city gets, the more it can keep tabs on our every move. Likewise, with connected home devices and digital assistants picking up our daily activities and queries, the potential for privacy breaches are endless.

Europe’s pioneering General Data Privacy Regulation (GDPR) is one of several attempts by governments to mitigate widespread shortfalls in customer data protection, for both companies and governments. Other countries, and even US states like California, have followed. But strict data regulations don’t come without their downsides: with more policies on the horizon across the globe, many worry future data-driven innovations and advances are at risk.

One potential solution: edge computing. Edge computing offers a way to reconcile the growing prevalence of IoT devices in our personal, public, and work lives with user privacy.

“Alexa, are you following the law?”

Voice assistants like Alexa and Siri continue to grow in popularity. In 2018, US smart voice assistant ownership grew 40 per cent, reaching 66.4 million—or 26 per cent of the US adult population.

The Alexa and Siri platforms may be “smart,” but the artificial intelligence capabilities that are needed to process our growing number of queries does not exist on the devices used to access them. Instead, a typical smartphone, for example, records what users say, transmits it to the server for processing, then beams back a result. “Alexa, can you play the latest hit by Justin Bieber?” travels endless miles to data centres and back before you get to actually enjoy the song in the comfort of your home. This is why Alexa and Siri generally don’t work offline (and why Google Home’s offline functionality is limited).

Much of the privacy concern brewing against smart home devices is actually not an issue with the devices themselves, but with the process of transmitting and storing data. That’s because data is at its most vulnerable in transit, and centralised databases of customer data serve as hotspots for hackers.

One answer? A truly smart device would be able to process all necessary data on its own. This would not only make the device more versatile, responsive, faster and more secure—it would also greatly aid in GDPR compliance.

To illustrate, let’s look at Amazon’s most recent Alexa-related privacy blunder. This July, Amazon accidentally sent a man 1,700 Alexa records that belonged to a different Alexa user. The kicker: these records were being delivered in response to the man’s GDPR data ownership request.

Now, imagine a world where customer data was stored and processed right on the device itself. Complying with a customer’s GDPR data request would be much easier—their data would be right on their device and nowhere else. And with new laws trending towards ever tighter enforcement of data accountability, building such protections directly into devices is key.

Data processing for the devices of tomorrow

These innovations could also apply to smart technologies well beyond the home, including smart cities. Many smart cities and cutting-edge public safety programs urgently need to orient themselves towards new regulations. GDPR, for example, restricts how potentially identifiable information can be stored, processed, and collected not just by corporations, but by municipalities as well.

Modern smart cities leverage large networks of cameras, cell phone and WIFI information, and various other data points to automate and optimise city services, traffic, and security.

Under GDPR, this data can be used without explicit consent only if it is anonymised and directly contributes to approved purposes. At the same time, performance of key functions like scanning video for missing persons, criminal activity, or traffic violations requires substantial processing power. If the data is processed outside the device, new privacy concerns come up. However, current device processors are limited in their ability to process information, and together, these limitations curtail crucial functionality.

To create a smart city, most of the computing power must be driven to the edge of the network – meaning the devices and sensors will be responsible for processing decisions locally, on their own, without a remote server. While there will always be at least some data sent to a server given the massive amount of information that needs to be processed, edge computing guarantees that only the most relevant, digested data will be sent there.  

GDRP- Inspiring a smarter future

A key solution to GDPR’s current limitations on innovation is edge computing. But edge computing alone is not enough. Intuitive edge processing with a dose of AI is key to making these devices truly smart.

AI-powered edge computing offers benefits far beyond GDPR compliance. These technologies are more scalable, dynamic and safer than what is generally used. GDPR is not impeding progress – it is inspiring a smarter, more sustainable future.

Orr Danon, CEO and Co-Founder, Hailo