top of page

Artificial Intelligence at the Edge

Where are your most powerful computers? Ask any IT manager that simple question and most probably they’ll point at the sky. Cloud computing is so dominant that it’s easy to forget that Satya Nadella raised an eyebrow or two when he ditched Microsoft’s mobile ambitions and bet the farm on Azure.

In hindsight, cloud computing is the logical conclusion of a 10-year trend that has seen processing power steadily drift away from on-premise data centers to one of the hosting giants including Amazon, Microsoft and, more lately, Google.

For most businesses, the argument is simple. Why shoulder the cost and the administrative burden when someone else can rent you an equivalent computing service that is less expensive, easier to scale and much more resilient?

So why is everyone so excited about Edge computing which appears to reverse this trend by redistributing processing power to computers and devices at the periphery of the network?

There are three forces at play here.

Firstly chip technology. Recent years have seen massive advances in processing power, especially on smartphones where the competition to deliver faster and more exciting experiences is greater than ever.

Then there’s the rise of machine learning and artificial intelligence. Much of the data that fuels machine learning is generated by the internet of things, which includes everything from driverless cars to air-pollution monitors.

But sending massive volumes of data from device to cloud and back again is inevitably constrained by network bandwidth, typically the ‘last mile’ which relies on local wireless and telco networks. If you’ve ever noticed a lag when entering data into an online spreadsheet, you’ll know what this means.

This is why everyone is so excited about the final piece in the puzzle, 5G. Take the example of driverless vehicles or assisted driving. Most of the technology that keeps you and your fellow road users safe is housed in the vehicle itself. But the full benefits of driverless technology are only achieved when vehicles communicate with one another. In theory this further streamlines traffic flows while increasing the safety of drivers, cyclists and pedestrians.

Healthcare is another sector that stands to benefit from this journey to the edge. Think of the millions of patients suffering from chronic diseases who could be monitored at home instead of visiting a crowded hospital every few weeks. Even within a hospital, the automation and verification of bedside health checks could reduce the burden on busy medical staff.

For consumers, more powerful chips and advances in machine learning are transforming the way they use smartphones and laptops. Advanced smartphone cameras have been trained to optimize landscape, portrait and food photography based on millions of online images.

They can even feed you prompts as you compose your image. Once the rule of thirds and the golden ratio were secrets known only to professional photographers. Now they’re embedded on most portable devices with a lens.

Of course, there are questions over privacy. More powerful chips and 5G networks enable organizations such as retailers, banks and car manufacturers to monitor customer behaviour in more detail than ever before. Tesla’s stratospheric valuation, for example, is partly based on the massive volumes labelled data returned by its vehicles, not just the vehicles themselves.

Finally, there’s the question of security. Each connected device is potentially vulnerable to hackers and other cyber criminals using it as a back door to the wider company network. Ironically security cameras are among the most compromised devices. But what happens when you can hack a driverless vehicle or a device monitoring the heart condition of a high-profile politician?

Right now there are far more ethical questions than answers. But as we've seen in the past, this won't stop the Googles, Apples and Teslas from pushing AI onto consumer devices from smart phones to smart cars. We've come a long way from the innocent days of virtual and cloud computing. Further still from email and printer servers. In every aspect of our lives from the workplace to our homes, the future lies at the edge.



Feel the connection?

bottom of page