How the Internet of Things (IoT) will be built is now pretty clear.
Of the vast sums of federal money soon to be pumped into the economy, a fair chunk will go to the cities. And the cities are going to instrument public infrastructure with thousands of sensors and actuators, which will connect to intermediate hubs, and on to the cloud in an optimal artificial-intelligence-driven (AI driven) exchange of information among various levels. According to the UN, 83% of Americans already lived in urban areas in 2018, so the rollout of IoT in the cities will determine much of our experience as citizens.
Vendors are eagerly lining up to supply the systems that will make all these services work, and Qualcomm is promoting a soup-to-nuts framework as well as specialty chips to nurture everything along.
To that end, the company held a two-day conference called Smart Cities Accelerate 2021 in San Diego (in person!) in late September.Speakers included the mayors of San Jose, CA, St. Louis, MO, and Miami, FL as well as a panel headlined by Earvin “Magic” Johnson, moderated by Jon Fortt of CNN, and attended by Magic’s investment partner, Jim Reynolds, as well as Qualcomm’s president, Cristiano Amon. Magic noted that a technology panel with three Black men on it was something new and used that moment to underline how equity must be built into the architecture of smart cities.
Needless to say, everyone there was gung ho on the idea of smart cities, in no small part because they will all tend to benefit from the adoption of smart-cities technology, which they are busy bringing into existence. But there’s a whole other angle to smart cities.
About three years ago, a BBC reporter named John Sudworth went to a police station in Guiyang, a city of nearly 5 million people in central China, and convinced the officers there to try an experiment. He posed for mug shots, and they gave him a few minutes lead.
Off he walked into the city and took various turns without being overly evasive. Why bother? At any given moment, he had no less than three cameras looking at him. The cops — who had uploaded his photos to their facial recognition AI and labeled him a person of interest — needed all of seven minutes to find him.
That was three years ago; they’re better at it now.The Chinese government says it can match every face in the country with a national identity card and a license plate. It really is a total surveillance state.
But, they reassured Sudworth that people living an exemplary life had nothing to worry about. If you do something wrong, they can use their AI to go back through a week’s worth of video and see all the places you were, with whom you met, and when you did what. But otherwise the data just sits there, eventually purged by new data.
Sudworth’s story was told me to over dinner the evening after the first day by Jeff Lorbeck, whose official title is Senior Vice President and General Manager, Connected Smart Systems at Qualcomm Technologies, Inc., which makes him essentially the guy tasked with shepherding Qualcomm’s IoT effort to success. Lorbeck lived in China for three years with his family and has traveled the back roads there.I had asked him what keeps him up at night and mentioned my own concerns about privacy in the IoT world. He told me that’s the thing that worries him the most.
In the United States, we kicked off the surveillance era by instrumenting ourselves with smartphones.Now, any Tesla owner is on the grid 100% of the time. And many cities have deployed traffic cams while businesses have put up security cams. You could say the your digital twin already exists and is only getting more detailed every day.So, what’s the big deal?
China already has nearly half a billion closed-circuit TV cams in its cities. London is famous for its camera density.U.S. cities already have tens of thousands of cameras. Counting private as well as public installations, by some counts the United States has a greater density of surveillance per capita than China.
Thus, we’ve entered the smart-cities era without much public conversation. In a country where most citizens barely study their own history, much less that of other countries, not many people paid attention to the experience of Sidewalk Labs in Toronto, Canada.
Sidewalk Labs, an urban developer owned by Alphabet (the holding company of Google), engaged with the City of Toronto to create a smart-city concept neighborhood. The city owned a 12-acre piece of land on Lake Ontario called Quayside.
The plan was to build an integrated urban environment with an advanced electronic infrastructure. Smart technology would govern everything from systems for managing access, heating, and cooling in buildings to facility usage statistics, human traffic management, lighting, and (of course) cameras.
The permitting process was impeccable, carried out responsibly, in the open, with full disclosure, comment periods, and questions answered. Arguably, the Canadian people knew what they were getting. But that didn’t prevent a veritable poo-storm when they finally figured out what it was. After vociferous protests about data governance and privacy, Sidewalk Labs canceled the project, disingenuously blaming the cancellation on Covid-19.
So, the sticky wicket isn’t the how. The technology is unfolding as we speak. But the why and whether.We haven’t stopped to ask ourselves what we will give up to gain the benefits of this even-more-intrusive level of integrated technology.
Perhaps the answer isn’t as difficult as we think. There seem to be two categories: industrial and consumer IoT.
Industrial IoT is an almost unmitigated good. A functioning water transport system that self-reports leakage saves both water and money. A system that optimizes lighting, electricity distribution, or a bus schedule is pure goodness. But when consumers get into the mix, the sword becomes double-edged.
One solution displayed by a partner at the show, a company called The Indoor Lab, made use of lidar (radar-like imaging done with lasers). The lidar images of people walking around at an indoor venue looked enough like people to distinguish them from, say, dogs, but there was no detail, just a scatter of points forming different shapes.
An application that was only counting people could use AI on these images and do a nearly perfect job, all without invading the privacy of the individuals in the field of view.
During a rapid-dating forum, I put the privacy question to each of the partners I spoke with at the tables.Most of them were software companies creating the glue to hold all these IoT elements together in a useful way. Each one admitted that blowback was a huge potential problem. Even these partners – who stand to benefit most from the development of this particular technology – recognized that once people understand what smart cities truly entail, they are likely to have a strong reaction.
Mattias Klein, CEO of Kognition, which does threat detection for private spaces using smart cameras, acknowledged that “Everything can be used for good or evil.”Some of the ways Kognition tries to mitigate that evil include keeping data decentralized, serving only private spaces (e.g., business premises), enforcing contract terms through license revocation, and limiting in software the number of cams deployable per license.
Tim Kustka, head of sales for FlightOps, a smart drone company based in Israel, hopefully asserted, “Over time, the public will see that the benefits outweigh the downside.” He also mentioned safeguards like geofencing and no-fly zones as mitigating privacy issues.
Amitabh Mathur, co-founder and COO of Truminds, admitted that his technology, which optimizes routes for transportation services, uses cams to count people. On a shuttle bus, for example, a cam does actual AI on each passenger, enabling full identification, but then throws away all that data and simply counts the person. Mathur recounted a personal circumstance in which he and nine family members took a vacation in wine country. They had rented an Airbnb, but when they got there, they found a camera mounted inside.Because several people in his party objected, they went home earlier than planned.
Bill Zierolf, a business development guy from Hitachi, called the consumer privacy issue “an almost impossible problem.” Hitachi tries to address such things by having a requirement that projects meet social, economic, and environmental as well as financial goals. He’s had project denied because they didn’t meet the “giveback” component.
In consumer IoT, then, there’s a ledger: good on one side; bad on the other.One could say this about any technology: guns protect; guns harm; the Internet teaches; kids waste hours on social media or playing stupid games; cameras catch criminals; cameras enable stalking.Much of consumer IoT cuts both ways.
On the good side, smart cities could provide educational tools that leverage central resources but have strong local components, launch drones that deliver medicines during critical time windows, catch criminals quickly, find lost or kidnapped children, and promote the welfare of citizens who have been left behind so far in the great digital build-out.
On the bad side, smart cities could destroy any expectation of privacy, allow dissidents to be rounded up arbitrarily, let stalkers who co-opt infrastructure track victims, and enable many other dystopian scenarios.
Given all that, the biggest problem facing smart cities will be selling them — with all their warts and imperfections fully disclosed — to the public.