Sidewalk Labs, the Candidate
By Sean McDonald and Bianca Wylie
Normally when we convey open-ended power over public spaces and policies, we do so to candidates through democratic election. In the Sidewalk Toronto case, through an open-ended request for proposal, that power may be offered to Sidewalk Labs via procurement.
But Sidewalk Labs is a company, not a candidate. When a candidate makes a promise on the campaign trail, we know that it’s marketing, and that we can’t rely on it. That’s one of the reasons we scrutinize the character of the candidate themself. So far, Sidewalk Labs has behaved like a candidate with a large and showy set of plans (overwhelmingly detailed in some places, extremely vague in others), rather than a vendor trying to convince us of its track record and proven ability to do the job.
So what if we treated Sidewalk Labs like a candidate? What if, in addition to reviewing their plan – their platform, if you will – we also reviewed the company’s character? What if we held them accountable for their performance and track record? And, what if we stopped ignoring the patron behind Sidewalk Labs’ public governance ambitions – Alphabet?
Sidewalk Labs
Sidewalk Labs pitches itself as an urban technology company, which means that its track record is comprised primarily of urban technology projects. Here are three of their highest profile projects and products, with a focus on community reception and some additional context.
Sidewalk Labs is four years old and is headquartered in New York, where it built its flagship project: LinkNYC – a product that replaced telephone booths with WiFi hotspots that display ads and messaging, and also collect and store users’ browsing history.
Sidewalk Labs was founded as a Google company before Google was reorganized into Alphabet. After Sidewalk Labs became an Alphabet company, Alphabet then co-invested with Sidewalk Labs in a company called Intersection, the company behind LinkNYC and, later, LinkLondon. The Link project has faced criticism from civil liberties groups based on the amount of sensitive data that it was allowed to collect, and the sensors and cameras installed in the units. Ultimately, Sidewalk Labs was part of a conglomerate that converted public phone booths into surveillant WiFi infrastructure, without any initial public audit, explanation, or oversight of data collection and use.
Sidewalk Labs also sells a product called Replica, which doesn’t disclose exactly where it gets its data, but has variably reported including data generated by Android phones (Android is a Google product), Google Apps, call detail records, third-party data brokers, and other sources of data to model mobility patterns for cities. Sidewalk Labs was awarded a 3.6 million USD dollar sole source contract with the state of Illinois Department of Transportation for Replica, and they are currently finalizing another deal to use Replica in Portland, Oregon. The company has committed not to use Replica in Toronto, in response to public concerns.
Replica, which has now been spun off into its own company, uses historical data as a way to inform policy makers about future movements – essentially, it’s ‘predictive policymaking’, which brings all the challenges of predictive modeling to public policy spaces. It has knowable short-term problems (black-box decision-making and bias being two of them) and the potential to create long-term dependence. If cities become reliant on new sources of data they can’t create themselves, how can they hold the vendor of that data to account?
These aren’t privacy problems, they’re fiscal and operational problems. This type of product, in the Sidewalk Labs’ Toronto Tomorrow plan, is called a ‘purposeful solution’ – a solution that they (Sidewalk Labs) say they have to create because no one else will or can. Purposeful solutions, of course, also ensure that Sidewalk Labs is the sole-source provider. In the Toronto plan, Sidewalk Labs would give these types of products to the City for free, demonstrating quite clearly that the organization plans to monetize its relationship with the City indirectly, limiting the normal leverage that’s available to negotiating parties.
Finally, of course, Sidewalk Labs has a track record in Toronto. That, maybe, is where we get a bit more into Sidewalk Labs’ character. We are, after all, in the middle of a campaign. Judging them solely by their actions, it’s hard to see them as good-faith negotiators. Sidewalk Labs exceeded the scope of Waterfront Toronto’s request for proposal (RFP), a snub to their “partner” now left at the feet of three levels of government for a solution. Even allowing for that, Sidewalk Labs’ approach to civic engagement has lacked transparency, has included willful misdirection about their business plan, as well as insincerity about the role of technology in their plans.
While the RFP did ask for a plan that would consider how to scale the project beyond the 12 acre Quayside site, Sidewalk Labs did not make their plans on how they would do this clear to the public during their public consultations. Sidewalk Labs has not talked much at all, and definitely not in detail, about what is possibly the largest and most important short-term part of their business plan: real estate. As an example, Sidewalk didn’t disclose their idea to claim a parcel of land separate from Quayside, named West Villiers island, for a Google headquarters, until the end of the process. If they’d been sharing their real estate plans in good faith, the leaked documents reported on by the Toronto Star wouldn’t have made any news.
Sidewalk Labs’ story about technology has been unclear, particularly in terms of how it will solve large, complex, urban challenges. When the project launched in 2017, Sidewalk talked about a neighbourhood built from the internet up, complete with rhetoric about the city as an operating system. Dan Doctoroff, chief executive of Sidewalk Labs, explained “The notion of the city as a platform — combining physical infrastructure, digital infrastructure, governance and then enabling people to create on top of it — is a completely different approach to city building.” Then, in June 2019 Doctoroff offered: “The data element and concerns around sensors has been a big focus of attention on this project and that misses the point. It was never actually about technology. Technology is just a means to help lower the cost of living, to improve health outcomes, the environment and to enhance safety.” Sidewalk Labs is an urban technology firm with a history of building invasive technologies. The idea that technology is incidental to Sidewalk Labs’ bid is disingenuous, it was one of the company’s core competitive advantages in the real estate procurement.
Sidewalk Labs’ response to criticism has been, largely, to deflect – to dismiss criticism as inevitable. For Sidewalk Labs the company, basic civic discourse, which must be grounded in the whole truth – not partial and conveniently timed truth, has proven difficult. Rather than engage the public in challenging conversations about the issues that are now threatening to derail the project (scope, for example) they hired and consulted with experts to secure local credibility, lobbied all levels of government, and released an almost impenetrable 1,500 page plan. While a certain amount of that may be inevitable in negotiations of this size, Sidewalk Labs seems to have assumed it could buy the governments’ approval – and regardless of whether it’s true, it’s not a particularly flattering statement, or the way to earn public trust.
Alphabet
In addition to Sidewalk Labs’ track record – and whatever you think of their corporate character – the truth is, it’s not just them. Not only does Sidewalk Labs have subsidiaries and investment interests, but, more importantly, Sidewalk Labs has an owner: Alphabet.
It’s probably useful to point out that Alphabet, the company, was founded during the second of three European Union antitrust investigations into Google. Google is currently under antitrust investigation in the United States, and is appealing a fine for anti-competitive practice in the EU. Larry Page and Sergei Brin, Google’s founders, now lead Alphabet – leaving Google to be run, under Alphabet’s oversight, by Sundar Pichai. According to its 2017 Annual Report, 86% of Alphabet’s annual revenue comes from Google. And, despite all the recent scandals, from YouTube to Project Dragonfly and from labour issues and a lawsuit, Alphabet’s 2019 earnings are up.
Doctoroff has gone to great pains, as recently as July 26th in a Reddit AMA, to assure the public that Sidewalk Labs is a wholly independent entity from Google. This is a common trope for Alphabet and its subsidiaries – they frame concerns around their relationship to Google, which they can change at whim, as alarmist speculation. Of course, Alphabet also has a history of merging its companies, and pieces of its companies, into Google – and then back out again. Proximity to Google’s technical proficiency is pulled out when it’s appealing, such as when questions of data security are raised, or when economic development claims are made, then walked back when it’s not.
In any case, Doctoroff’s underlying assertion is 100% true, today. Google and Sidewalk Labs have been explicitly engineered for the appearance of separateness, though they are owned by the same parent company, which is run by Google’s founders, draw on the same digital infrastructure, and share a common interest in finding new ways to collect and use data. Still, for argument’s sake – let’s assume good-faith. Let’s assume that Google’s interests are currently separated from Sidewalk Labs’ – so, let’s only look at how Alphabet companies have done at maintaining promises to remain separate from each other, keep data private, and to support public infrastructure.
First, let’s start with Nest, a smart home thermostat, security camera, and alarm company first acquired by Google in 2014 for $3.2 billion USD, amidst promises to users to remain totally separate from the rest of the company. Nest was one of the first companies managed separately when Google spun Alphabet up – and its independence was a key privacy selling point. In 2018, just over 4 years later, Google re-absorbed Nest, merging their data infrastructure and integrating product lines. Parent companies, it turns out, can use their company structure to invalidate the promises made by their subsidiaries. Alphabet can choose to merge any of its companies, at any time, with Google.
Alphabet made almost exactly the same decision with portions of DeepMind, an artificial intelligence company it acquired in 2014. DeepMind had a controversial data sharing partnership with the United Kingdom’s National Health Service – which encouraged enough outcry to require an independent auditor. In late 2018, Alphabet decided to re-absorb portions of DeepMind into Google – leaving other parts independent and in-place. Alphabet can also merge portions of its companies’ with Google. As Julia Powles has been tracking, this doesn’t even begin to get into whether or not the stated policy outcomes related to the technology are going to happen.
Let’s assume, however, that the three levels of Canadian government figure out a way to ensure that Sidewalk Labs’, as a company or as a set of digital assets, can be ring-fenced from Google. What’s Alphabet’s track record with public infrastructure?
Google Fibre, confusingly, is not a Google company – it’s a standalone part of the Alphabet Access Division. Google Fibre offers internet connectivity in a wide range of American cities, but notably abandoned its investment in Louisville, after a high-profile engineering failure. As the company explained: “When we launched Fiber service in Louisville in October 2017, we noted at the time that it was the fastest we’ve ever moved from construction announcement to signing up customers. That’s because we were trialing a lot of things in Louisville, including a different type of construction method — namely, placing fiber in much shallower trenches than we’ve done elsewhere. Innovating means learning, and sometimes, unfortunately, you learn by failing.”
Louisville created new rules for Google Fibre’s technology, paid to manage related legal challenges from telecommunications companies, then ended up as a failed experiment. When testing minimum viable product (MVP) techniques, and taking less precautions than normal in order to outpace competitors, the company sees it as learning. But another way to think about this is as an experiment that didn’t factor in resident needs, should it fail. Instead of sticking it out, or rebuilding the infrastructure, Alphabet chose to abandon the project – giving residents 2 months of free service to cushion the blow. Alphabet can choose to leave a project without the need to provide a viable alternative for people who depend on them.
Similarly, Project Loon is an Alphabet company that provides balloon-based backhaul internet connectivity to several countries, whose privacy policy directly shares data with Google’s advertising service – raising questions about what independence really means for an Alphabet company.
Ultimately, Sidewalk Labs can claim independence from Alphabet and be, for now, correct – but Alphabet has proven that it will restructure its companies, combining them with Google, regardless of promises the company made beforehand. So, Sidewalk Labs’ assurances of its independence from Alphabet or Google, while factually accurate today, may change after the critical part of the deal is done. They have before.
So how to assess Sidewalk Labs the candidate? Well. It has a history of building and deploying invasive technology infrastructure, without much detail on its sources, its subsequent uses, or how it actually solves publicly important problems. In a campaign, of course, we’d have election financing laws for transparency, an understanding of the candidate’s recent professional history for level, and a platform that outlines what kind of world they’re building and how. Here, we don’t have a clear picture of the revenue, tax, corporate structure, investment, or data architecture information the public needs to evaluate Sidewalk Labs’ independence or credibility.
Without this, we can look to some of the recent and longer-term history we’ve laid out here to consider how this candidate’s record lines up alongside the platform it has submitted. The plan can and should be studied and assessed on its own merits, in includes several good ideas, but they’re ideas that aren’t necessarily new and don’t require Sidewalk Labs. It’s the character of the candidate that determines if and how these goals might or might not be achieved, as well of some of the additional issues to consider when thinking about whether or not to take this proposed deal any further.
The current round of negotiations has just added an October 31st off-ramp for either Sidewalk Labs or Waterfront Toronto to walk away from the deal should Sidewalk Labs not address some of the key issues identified with the current proposal. There is an opportunity, should this deal end on that date, to take the plan and still make good use of some of it. As with any big election, sometimes platforms outlast candidates. What’s important is finding leadership that can be trusted to help build out a resident-led vision for the future of Toronto’s waterfront.
Image: Karl-Ludwig Poggemann
