Inside Google’s plan to build a smart neighborhood in Toronto
Sidewalk Labs has a grand vision for the Eastern Waterfront.
On the Sidewalk Labs website is a 200-page document explaining its vision for a smart neighborhood in Toronto. It's packed with illustrations that show a warm, idyllic community full of grassy parks, modular buildings and underground tunnels with delivery robots and internet cabling inside. The text describes "a truly complete community" that's free of cars and committed to reducing its carbon footprint. Underpinning everything is a network of sensors that can monitor noise, traffic and pollution, collecting the troves of data required to understand and improve the city's design.
Flipping through the pages, it's easy to see how the company -- an offshoot of Google parent Alphabet -- was chosen to revitalize the Lake Ontario waterfront. The lengthy pitch document, however, is just a taste of what the area might become. It's a dreamy but meticulously thought-out mood board summarizing what Sidewalk Labs has been pondering for the past two years. Reading it cover to cover, you can get lost in the scale and ambition of such a project. Most companies would struggle to execute just one aspect of the plan: autonomous transit, for instance, or buildings that can be quickly and cheaply repurposed depending on the time of day or needs of the city.
Sidewalk Labs, however, wants to do it all.
The project started with an email sent by Eric Schmidt, Google's then executive chairman, to Dan Doctoroff in 2014. The subject line read, "The City of the Future." Doctoroff was the head of Bloomberg LP, an umbrella company for its terminal business, news wire service and journalistic ventures. Previously, he had worked with Michael Bloomberg as New York's deputy mayor for economic development and rebuilding. Schmidt wanted to know if Doctoroff would meet Google co-founders Larry Page and Sergey Brin and listen to their thoughts on smart cities.
At the time, Google was working on its smiley face self-driving car. Behind the scenes, however, the company was having broader conversations about technology and how it could be used to improve urban life. Page and Brin were enthusiastic but knew they needed a specialist, or some kind of leader, who understood cities as much as technology. Their conversations with Doctoroff were fruitful, and in the summer of 2015, Sidewalk Labs was announced to the world. "Every time I talk with Dan I feel an amazing sense of opportunity because of all the ways technology can help transform cities to be more livable, flexible and vibrant," Page said in a Google+ post.
Two months later, Google announced Alphabet, a new business structure that made Google a subsidiary and its many moonshot projects, including delivery drones and life extension, standalone businesses. Sidewalk Labs was "a new company" from the start, but this corporate shuffle solidified its position as a long-term, experimental bet.
Doctoroff and his team spent the next two years studying more than 100 urban initiatives. This served as preparation for the Waterfront Toronto project, which launched in March 2017. The City of Toronto, with support from the Canadian government, was looking for a partner to rebuild and revitalize roughly 750 acres of land along the Eastern Waterfront. The first phase, or pilot, would be Quayside, a 12-acre site close to the central business district. Several local and international firms submitted proposals, including Sidewalk Labs. On October 17th, 2017, the Alphabet-owned offshoot announced that it had won the bid and given the project a new name: Sidewalk Toronto.
So-called smart cities have been tried before with varying levels of success. One of the most ambitious is Songdo, a tech metropolis built 40 miles west of Seoul, the South Korean capital. In 2000, it was a marshy patch of tidal flats; now it's a connected city dominated by glassy skyscrapers and a New York-inspired Central Park. Cameras are scattered across bridges, highways and narrow back alleys, providing an endless stream of video to a human team stationed in Songdo's G-Tower. They monitor traffic for accidents and congestion as well as natural disasters, crime and public facilities that might require repairs.
Residents have touchscreen panels that allow them to control the temperature and lighting in their homes as well as the timing of deliveries. A citywide pneumatic refuse system sucks garbage below the surface and into a remote sorting center, eliminating the need for dirty garbage trucks. Songdo can feel a bit quiet and sterile, however, with rent prices that exceed the budget of the average household. It was meant to be a hub for international businesses -- Incheon International Airport is a short drive away -- but adoption has been slow. French newspaper Le Monde described the city as a "ghetto for the affluent" last year.
"Most smart cities projects are in a very early stage."
Singapore is using a network of sensors to understand people's energy usage and waste production. It's also considering a mandatory GPS system that tracks every car on the road in real time. But there are privacy and security concerns born out of a government that is often described as authoritarian. Barcelona, meanwhile, has deployed an army of sensors to track air quality, free parking spaces and the amount of trash in public bins. It's also dabbling with a superblock concept that limits the speed and volume of traffic inside small metropolitan areas, prioritizing the movement of cyclists and pedestrians instead.
These projects are small, however, compared to Neom, a planned city on the Red Sea coast that will be powered by wind turbines and vast fields of solar panels, announced last October by Saudi Crown Prince Mohammed bin Salman. According to an ostentatious promotional website, the streets will be occupied by "automated, 100-percent green transport systems," including passenger drones. Food will be supplied by giant vertical urban farms while residents enjoy "the world's largest garden," endless natural parkland and record-breaking theme parks.
Many are skeptical. "Most smart cities projects are in a very early stage," Cesar Cerrudo, founder of the not-for-profit Securing Smart Cities, said. "If you do your research, you'll find there are a lot of projects but there isn't much that's concrete. It's a lot of publicity, a lot of marketing."
Sidewalk Labs believes its dual expertise in technology and urban design makes it different. The company is based in New York rather than Silicon Valley to distill a sense of city living in its employees. That DNA also shaped its bid and, consequently, the ideas that it's pushing forward for Quayside. The buildings, for instance, will use a modular design that's cheaper and faster to build. Some of them will utilize Loft, a minimalistic interior that means they can be quickly repurposed. In this scenario, a parking lot could be converted into an office as more people start to embrace electric and autonomous ride-sharing options.
Buildings will be made from eco-friendly buildings materials, including tall timber skeletons and mycelium insulation, and powered by renewable energy sources, including roof and wall-mounted solar panels. They'll be warmed and cooled by a thermal grid that leverages waste heat from sewers and buildings as well as geothermal sources and nearby lakes. Homes will also be LEED (Leadership in Energy and Environmental Design) certified and meet the energy-efficient Passive House standard developed by two professors in Germany (where it's known as Passivhaus) in the early 1990s.
"By far the most important issue is affordability."
Sidewalk is committed to affordability and avoiding Songdo's reputation as a half-empty city for the rich. It's banking on a blend of simple and old-fashioned pricing models, including social tenants, private renters, subsidized renters and full-fledged homeowners. "A partial homeownership program might be a perfect fit for a family looking to settle down, whereas a retiree on a fixed income may require a rental subsidy," the company explained. "Sidewalk proposes to make Quayside a living laboratory for housing policy innovation that delivers a mixed- occupancy community that mirrors Toronto's socioeconomic diversity."
Doctoroff pushed this point at a community town hall meeting last November. "By far the most important issue is all about affordability," he said. "We certainly see everyday people, middle class, lower-middle and lower-income people, being priced out of areas, suffering from the incapacity to get opportunities because distances are too great or the costs are too great. If we don't fundamentally address that issue, the fabric of society and the whole notion of this being an inclusive community begins to fray."
Sidewalk Labs is serious about ditching cars too. The team wants the Eastern Waterfront to be the first district in Toronto where only shared and self-driving vehicles are allowed. Non-emergency vehicles will be banned from "a large portion" of the neighborhood, giving space back to pedestrians and cyclists. A "transition zone" will exist for people who need to travel beyond Quayside, but the idea is to promote a walking, cycling and public transit culture in the center.
The absence of private cars should make the roads feel quieter and safer. It will also eliminate the need for curbside parking, freeing up crucial space for sidewalks and stores. "It's all about getting rid of stationary cars," Alan Penn, a professor in Architectural and Urban Computing at UCL, said. "That can transform what somewhere feels like."
Sidewalk plans to extend Toronto's existing bike-share scheme into Quayside. It's also considering an LED system that can create, widen and narrow temporary bike lanes on the road. The latter is experimental but perhaps unnecessary. Michael Seth Wexler, an urban designer at cycling consultancy Copenhagenize, said permanent, protected cycleways will be critical if the city wants to get more people on two wheels. If they disappear or they're few and far between, people won't be able to rely on them -- and casual cyclists will default to other forms of transportation.
"It's all about getting rid of stationary cars. That can transform what somewhere feels like."
"To know that there's always going to be a protected path that you can take down the main street," Wexler explained, "before any modern technology is implemented, that's a baseline that will create a sense of reliability in the transport system."
Above all, people want to feel safe. "They won't feel safe unless they have a separation from other modes of transport," Wexler added. "You don't want to be cycling slowly with a tram zooming right by you and then you have to duck out of the way of a transport truck or even an automated vehicle." It's worrying, he said, that none of the illustrations in the Sidewalk Labs pitch document show a protected cycleway.
Wexler welcomes technology, however, and how it could make cycling a more efficient and attractive option in Quayside. Sidewalk Labs will pilot an "adaptive traffic light" concept that can detect and prioritize cyclists at busy intersections. It would be similar to the green-wave system in Copenhagen, which ensures that cyclists traveling at 20KMH hit green lights all the way into the city center. The company has some new ideas, too, such as automatic, retractable canopies and heated bike paths that melt snow in the winter.
Technologists vs. urbanists
Doctoroff was brought onboard to bring urban planning and design expertise to Google's traditionally technological background. Blending these two worlds hasn't been easy though. Developers will often use A/B testing, for instance -- two versions of the same software, quietly distributed to different users -- to see which solution works best. In a city, however, that's often not an option. You can't A/B test a traffic light system if there's a chance one of them will cause accidents on the road. Similarly, "fail fast, fail often" is a bad ethos when lives are at risk.
Urbanists act more deliberately but can often see solutions as unchangeable. When a home is built or a cycle path is put in place, they see it as a semipermanent development. Sidewalk Labs found it "difficult" to balance these two cultures in its first year of operations. "It led to some conflicts," said Rit Aggarwala, chief policy officer at Sidewalk Labs. "It led to a bunch of misunderstandings where it wasn't an intentional conflict. We just realized we weren't speaking the same language with each other."
The urbanists had to be more explicit about the risks they were identifying, and the technologists had to be more attuned to where something could go wrong. Similarly, the developers needed to retain their enthusiasm and speed while protecting the values that urbanists believe are important.
Sidewalk Labs is envisioning Quayside as a series of layers. At the bottom is a network of tunnels, or utility channels, which serves as the city's near-invisible infrastructure. Above is the public realm, or street level, which serves as a foundation for its mobility and building concepts. At the highest level is the digital layer, which combines a network of sensors, a detailed map of the neighborhood, simulation software and a platform where citizens can log in and manage their public and private data.
Google makes its money by tracking internet users and serving them highly targeted ads. It's the revenue engine for Alphabet that, in turn, allows Sidewalk Labs to operate so freely. As a result, people are worried about Quayside and how residents' data will be treated. Will Sidewalk Labs have full access to the information collected by its sensors? How will data be shared with the company's partners? Will citizens be able to opt out of certain programs, and if so, how will it affect their quality of life in the city?
People are worried about Quayside and how residents' data will be treated.
Aggarwala is promising a privacy-by-design approach. That means limiting data capture to the "bare minimum" throughout the city. The company's internal Sense Lab, for instance, is developing a camera system that strips surveillance footage down to a series of faint outlines. "We don't need an image of you," Aggarwala said. "What we need is your outline, because then the computer can tell, 'Oh, that's a human. That's a person walking.' If all I do is outline your body and there's no face, no color, no nothing, then there's no way I can identify you. I've eliminated the privacy issue, but I've accomplished the goal."
Not everyone is convinced. To make a service smarter, or more user-friendly, you generally need more information. Let's say someone clambers into an elevator; with a motion sensor you can detect their presence and light up a control panel. If you know who they are, however, you can also play their favorite song or take them to their hotel room automatically. "They can anonymize the information," Cerrudo explained, "but in the end, the less they anonymize it, the more functionality they can give to the end user."
Part of the solution is a simulation platform called Model. It will be "metro area scale" and cover the movement of every Quayside resident who might wander through its virtual net. Sidewalk Labs will use this data to test possible changes to roadway pricing, ride-sharing and its multiuse buildings. It will also accept data collected by the Sense Lab team to simulate what will happen in the next five, 15 and 30 minutes. Over time, the company will use the platform to test long-term changes to water, energy and other public infrastructure. It will also grow more accurate and sophisticated with use, to the point where someone could ask, "Where were the 20 people on this bus really starting from, and where were they really going?"
"It's a tool that doesn't exist," Aggarwala said. "We need to build it because the state-of-the-art equivalents right now are very expensive, very slow and very coarse in terms of their analysis."
It seems inevitable, though, that Sidewalk Labs will need to change or upgrade one of its services in the future. When that happens, the company will have to survey Quayside residents or ask them for approval, just like developers do now with user agreements in the App Store. Cerrudo suspects the company will introduce an "end-citizen" agreement that includes automatic approval for all technological changes in the city. Otherwise, "the challenge is to find a way to easily ask people for feedback and get an answer," Cerrudo said, "and then decide what to do."
Consultations will be a political and democratic problem as much as a privacy and security issue. Decisions will need to be backed by Quayside citizens as well as local and national government. "It could be very bureaucratic," Cerrudo said. Sidewalk Labs, however, thinks it's a question of trust. If its policies are transparent enough and there are robust safeguards to protect user data, residents will slowly give Doctoroff and his team the benefit of the doubt. "If not, then frankly we've made a mistake, because we have to earn that trust," Aggarwala said.
"There should be a whole chapter on privacy and security. There shouldn't just be one reference."
To do so, Sidewalk Labs will be asking for help from independent cybersecurity experts. Ann Cavoukian, a former information and privacy commissioner for Ontario and now executive director of the Privacy and Big Data Institute at Ryerson University, has agreed to serve on the company's advisory board. It's a positive step, though Cerrudo is still concerned about the company's commitment to security. In the pitch document, for instance, "I looked for 'security' and there are just two mentions of security, on two pages," he said. "There should be a whole chapter on privacy and security. There shouldn't just be one reference."
Trust will also hinge on the people that Sidewalk Labs is able to attract to the area. The company wants the community to be multigenerational, mixed-income and diverse. If all of its residents are Google employees or Silicon Valley developer types, the project may be seen by many as a failure. That doesn't mean Quayside will be for everyone, however. A major opportunity, Aggarwala said, will be attracting people who are open to urban innovation.
The company is thinking about a Smart Chute system, for instance, that tracks how much waste people are throwing away. It would then be possible to charge residents per bag, promoting waste reduction and generating cash for city refurbishments. Such a move would also help Quayside to reduce its environmental impact and the amount of traffic running through its underground tunnels. To be effective, though, Sidewalk Labs needs residents who can understand and embrace the long-term benefits. If a path is widened or a canopy is moved, citizens need to trust that it's for the greater good.
"My strong belief is that we're going to find it's an attitude that cuts across age, gender, income, background -- all sorts of things," Aggarwala said.
Waterfront Toronto has recognized the importance of privacy and data governance too. In a memo, Kristina Verner, vice president of innovation, sustainability and prosperity at Waterfront Toronto, said Cavoukian's help and privacy by design were "not sufficient to ensure legal compliance with relevant legislation." The organization will, therefore, create an independent Digital Strategy Advisory Panel to assist on the project. Members will be plucked from the academic, legal and civic technology communities and help Waterfront Toronto shape its digital-governance requirements.
Sidewalk Labs is an offshoot of Alphabet, but it's also experimenting with a subsidiary structure of its own. That's unusual; most of Google's acquisitions are quickly absorbed into one of its existing businesses. In June 2015, Doctoroff and his team led a consortium in the acquisition of Control Group, a New York-based technology and design consultancy firm, and advertising agency Titan. The pair were combined and rebranded as Intersection, a standalone company that has since turned thousands of old pay phones into free, gigabit-speed WiFi kiosks.
On February 1st, 2018, Sidewalk Labs announced another spin-off called Coord. The company is developing a platform with APIs that relate to road tolls, curbs and parking. It's a small but important piece in the development of a smart city. With a custom surveyor app, for instance, Coord employees can quickly photograph parking signs and other curbside information, digitizing road regulations in minutes. Developers can then access that information as an API and use it to improve their navigation apps and software. If you run a delivery truck business, for example, you need to know where it's OK to park and unload. Similarly, taxi drivers want to find the best places to pick up and drop off customers.
In the future, Coord will help self-driving vehicles adjust their routes in real time, preventing congestion, double-parking and unwanted toll fees. Google Maps has already integrated Coord so that drivers can find available parking near their destination. "By giving people full insight into all of their trip choices -- and seamless door-to-door service -- tools like this one can lead to better mobility outcomes for cities," Stephen Smyth, CEO of Coord, said in a blog post.
It's not clear if or how Coord and Intersection will work on Sidewalk Toronto. But a distribution of expertise could help Doctoroff's company to focus and solve specific problems.
In many respects, Sidewalk Labs is building Quayside from scratch. The coastline area already exists, but the company is imagining a complete revamp that includes roads, buildings and public areas. It's an opportunity for the company -- a chance to implement large, experimental infrastructure from the outset -- but also a great challenge. The more options that you have, the harder it is to make decisions and prove that a single component or change is having a positive impact. Penn calls it a "combinatorial explosion." "People often think that a clean slate is an easier thing to handle in design," he said. "In fact it's not. It's much, much harder."
Most designers circumvent this problem by creating artificial constraints. They'll look at real-world examples and choose a small selection as possible solutions. Or they'll settle on a particular style that drastically reduces the possibilities in an urban area. Penn said, "You can say, 'Well, actually, I'm going to build a city in a big green field.' But how do you decide anything? Where do you start? So the first thing you do is invent constraints or you look at what constraints there are in the world. That narrows the possibilities down to the point at which you have a starting point."
It's also inaccurate to call Quayside a clean slate, because it sits within the larger context of Toronto. It's part of an old, established city that has its own culture and expectations. Many citizens, for instance, use cars to run errands and get to work. If the Eastern Waterfront has a complete ban on private vehicles, that will make it a difficult place to visit -- a futuristic but isolated community. That's why the company is thinking about a transition zone. Aggarwala compares it to Canary Wharf, a business district in London that was originally served by the DLR train line exclusively. It felt remote until an underground station was added as part of the 1999 Jubilee Line extension.
"All of a sudden it felt like, 'Oh, that's part of the city because that's part of what the rest of the city does.'"
With Quayside, Sidewalk Labs wants to build a neighborhood with a strong sense of community. The company envisions "a next-gen bazaar" where residents can sell their "tech-enabled" wares or teach others new skills. The dense mixture of homes, stores and public facilities will encourage people to walk around and interact with one another. Community is an abstract value, however, that's difficult to measure. It's tied to mental health and how people feel about a place more than their physical actions.
How will Sidewalk Labs know, then, when its urban project has been successful? "That's a really good question," Aggarwala said. "Those things do not lend themselves to quantifiable ... to measurement." Techniques exist to measure relationships, but they're often crude and lack nuance. At the end of a technology conference, for instance, you can count the number of business cards someone has come away with; however, that doesn't explain how many of those interactions were deep or meaningful. You could ask how many times those people contacted one another again, but even that is a shallow assessment. One life-changing relationship, for instance, is better than a dozen emails that never went anywhere.
"We'll have to experiment, and we'll know it when we see it," Aggarwala said. As a starting point, the company has identified 25 "success metrics" that relate to someone's quality of life. These include the cost of rent and transportation as well as carbon emissions, job opportunities, park access, civic participation and time spent commuting. All of them will be tracked using sensors and the company's Model platform.
"The idea that this somehow can't make money or can only make money if Sidewalk Labs is capturing and monetizing private data -- I don't buy that at all."
Many wonder how Sidewalk Labs will make money on Quayside. The company's insistence on privacy by design suggests it won't be harvesting data like its sister company Google. Aggarwala said it won't be an advertising-based model, so there won't be giant billboards beaming targeted ads like in Blade Runner 2049. When pressed, he said, "It would be inaccurate for me to say that we're 100 percent confident that we know exactly what the business model for us, through this project, is."
The company has some time to figure it out. It will spend most of this year developing ideas, listening to the public and crafting a Master Innovation and Development Plan. That final document will need to be approved by Waterfront Toronto, the public and a variety of government bodies before a single brick can be laid. Aggarwala acknowledges that his company, and the project, will need to generate revenue. If it struggles to break even or becomes a piece of Alphabet philanthropy, it won't inspire other cities to do the same.
There will be opportunities, though, for Sidewalk Labs to charge a fee for its services. A city, after all, contains landlords, transport operators and utility companies; Doctoroff's company could bill any of these for access to its technology. "The idea that this somehow can't make money or can only make money if Sidewalk Labs is capturing and monetizing private data -- I don't buy that at all," Aggarwala said.
Quayside will serve as a test bed for many Alphabet technologies. Once they're perfected, Sidewalk could profit from selling the individual pieces -- the blueprint behind its modular buildings, for instance, or its simulation platform -- to other cities. Revenue, then, wouldn't come from advertising but making Quayside a living, breathing advertisement for other city planners. Success, Cerrudo said, would be "great for PR and marketing, and enable them to make money in the future when the technology is widely adopted."
Doctoroff knows how important it is to win the public's support. On November 1st, 2017, he participated in a community town hall with Will Fleissig, CEO of Waterfront Toronto, which was open to the public and streamed online. Fleissig was keen to stress that nothing about the project is set in stone. "It starts tonight," he said. "We're having this discussion, and we're listening. There is no plan in place. We can actually co-create by bringing together a lot of different partners, and together, we'll figure out what should happen on the Waterfront."
The company has since published a Public Engagement Plan that specifies how it will discuss and consult with Toronto residents. It includes a series of public talks, roundtable meetings and a pavilion where people can learn more about the project, participate in workshops and tinker with interactive exhibits. There will also be design jams, a Fellows Program for 19-to-24-year-olds and a 36-member Reference Panel that will meet throughout the year and take a closer look at Sidewalk Labs' work.
Later this year, the company will conduct a series of pilots to show the public some of its thinking. Aggarwala is staying tight-lipped about the details, but one is likely to including self-driving vehicles. That could be Waymo, Alphabet's autonomous car spin-off, or another industry specialist. "There has, to my knowledge anyway, never been an opportunity for somebody in Toronto to ride an autonomous vehicle," Aggarwala said. "We think we can help address that."
Wexler thinks this is the right approach. Pilots are cheaper than a full rollout and allow companies to get valuable data that can be used to refine the final idea. It beats the antiquated model, anyway, of a token-gesture consultation and a shiny but ultimately unfit-for-purpose skyscraper. "Do pilot projects," Wexler said. "Do pilot projects and collect data. And then use that data to understand if what you're doing works, and then tweak it."
At any moment, the city of Toronto could sever its partnership with Sidewalk Labs. The bid the company won is for a planning exercise that will run for at least a year. Doctoroff and his team are putting $50 million into the project, so clearly they're committed to building a smart neighborhood. But Waterfront Toronto, and the people it represents, can back out at any time. "At the end of the year, if you don't like it, if Waterfront Toronto's board doesn't like it, if the elected officials don't like it, they can tell us to go bye," Doctoroff said at the town hall meeting.
It's a big gamble, but one that Sidewalk Labs has been preparing for since 2015. The company has ideas, and it's confident that with proper consultation it can whip up a development plan that's hard for the city to refuse. "We're going to do our best possible work," Aggarwala said, "and I think they will find it compelling."