Council Post: The Promise Of Smart Cities Finally Coming True

Lee-Lean Shu, CEO, GSI Technology.

getty

​Cities are changing in ways most people are only now starting to notice. To put it simply, they’re getting a lot smarter.

Sure, we’ve heard about smart cities for over a decade, but those early promises and predictions of a fully connected urban life have never really come to fruition. Now, however, things are starting to come into focus.

But before we get there, let’s take a step back.

The first wave of IoT-powered “smart” products to hit the market were tailored for the home: things like smart locks, smart bulbs, smart thermostats and smart smoke detectors. Sure, as individual products they were pretty cool. But the overarching benefits of greater efficiency, safety and cost savings never materialized. That’s probably why the bigger vision of a fully integrated smart home never really caught on.

Sure, you could build your own integrations with Amazon Alexa, Google Home and a pile of compatible third-party devices, but that was limited to DIY hobbyists and high-end installations. As a result, the smart home never went mainstream. Commercial building service providers did a bit better. Integrators connected lighting, HVAC and power systems, but even that was a narrower win than the industry once imagined.

So what’s different now? The answer is edge convergence. The sensors, low-cost microcontrollers and ubiquitous connectivity that excited everyone about IoT in the first place have now been joined by cloud-based AI and cheap remote cameras.

Put all those pieces together with AI and you get a level of automation that simply wasn’t economical a decade ago. “Smart cities” can now use cloud-based AI to spot patterns in real time, and that capability is being put to work on public safety.

The Huge Payoff Of Smarter Cities

This evolution has quietly produced a market that’s already at a trillion dollars and is rapidly growing. According to Grand View Research, the global smart cities market is on track to increase nearly 30% a year and reach $3.76 trillion by 2030. This growth is being driven by government investment in emerging technologies like AI, cybersecurity and IoT.

The government investment is happening because, these days, more and more municipalities are coming under pressure to achieve sustainability benefits, as well as the need for increased efficiency in city planning as city growth pressure mounts. Municipalities are turning to smart technology to achieve this.

To be sure, it’s not been smooth. For instance, a lot of cities put cameras in place and collected terabytes of data as the cameras sat on traffic poles or monitored public spaces, but then didn’t efficiently use all that data. So much of it went nowhere. There were also privacy concerns that stalled deployment.

That’s changing now. With anonymizing techniques like face obscuring, combined with AI vision search and the kind of contextual awareness that vision-language models offer, cities can identify incidents and threats without compromising privacy. They can spot infrastructure failures, criminal activity and traffic incidents as they happen, without invading privacy. They can monitor human interactions for safety in public institutions like schools, senior living centers and prisons.

Unlike those predictions of the past, this time the smart trend is really showing benefits. For example, Mordor Intelligence notes in its Smart Cities Market Report that Singapore’s Smart Nation AI initiative has trimmed average commute times by 11 minutes thanks to a 92% accurate bus-crowding prediction model. Meanwhile, New York City’s FloodNet uses ultrasonic range-finding sensors and edge processing to calculate flood depth in real time, helping city agencies make more informed decisions during times of heavy weather. Taiwan’s Hsinchu County is implementing the Smart Eye Guardian system to automate the monitoring of public-safety events, including hazards, fires, violent activity, traffic incidents and suspicious behavior.

Compute Closer To The Action

Adoption acceleration will come from moving processing to the edge. This is being enabled by newer architectures like compute-in-memory that will increase performance and reduce costs. Compute-in-memory is a completely different kind of architectural approach that can dramatically increase inference capability as well as drastically improve energy efficiency by doing much of the AI processing at the edge where the action is.

The benefits for smart cities are straightforward: lower cost, less internet dependency and latency and faster identification of problems. And since these systems are constantly evaluating new images in real time, that speed, or time to first token (TTFT), matters a lot. In edge AI cases like smart infrastructure, responsiveness is everything.

As AI compute moves to the edge, the lower power and lower latency processing will also enable more use cases. Being able to process the multi-sensor data at the police station, energy station or grid interchange will likely invoke Jevons paradox. Namely, the falling cost of these capabilities, both in terms of environmental impact but also in absolute dollars, will increase demand, thus providing wider areas of advantage.

With improved secure automation and convergence, infrastructure itself is getting better. Power grids can, in real time, draw energy across multiple sources in the most efficient manner. Traffic signals can adjust on the fly based on actual congestion. Transit systems can give riders live updates. Taken altogether, these improvements are reducing energy costs, making power grids more resilient and influencing more people to use public transit.

Gone will be the days of a person surveilling walls of video monitors for eight hours waiting for an incident, and then somehow missing it when it actually occurs. Smart cities can now find problems automatically and highlight to the human observer an incident that requires further evaluation. Emergency services can also be notified in parallel, resulting in faster responses. The real promise here isn’t automation for its own sake. It’s safer, better functioning smart cities that finally live up to their billing.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?