Case Study: The Sidewalk Labs Toronto Data Trust

"People's data should not be treated like a natural resource to be extracted. It should be treated as something that people have a right to control." — Jim Balsillie, former co-CEO of BlackBerry, testimony to Canadian Parliament, 2019

Overview

In October 2017, Waterfront Toronto — a tri-government agency responsible for revitalizing Toronto's waterfront — announced a partnership with Sidewalk Labs, a subsidiary of Alphabet (Google's parent company). The plan was ambitious: transform Quayside, a 12-acre parcel of underused industrial land on Toronto's eastern waterfront, into a "neighborhood built from the internet up." Smart sensors would monitor traffic flows, air quality, noise levels, pedestrian movement, and energy usage. Data would optimize everything from garbage collection to public transit schedules. It was to be the model smart city of the twenty-first century.

At the center of the proposal was a novel governance mechanism: a civic data trust that would manage the data collected across the neighborhood. The trust was meant to resolve the ownership question — ensuring that the data generated by the community belonged to the community, not to Google. What followed over the next three years was one of the most significant public debates in the history of data governance, culminating in the project's cancellation in May 2020.

This case study examines what the Sidewalk Labs data trust proposed, why it provoked such intense resistance, and what the episode reveals about the promises and limits of data trusts as governance mechanisms.

Skills Applied: - Evaluating data trust design against the principles described in Section 3.3.1 - Analyzing stakeholder interests using the stakeholder map from Section 3.1.2 - Assessing how theories of data ownership inform real governance disputes - Identifying the gap between governance design and governance legitimacy

The Situation

The Vision

Sidewalk Labs, led by former New York City deputy mayor Dan Doctoroff, presented an extraordinary vision. Quayside would feature:

• Adaptive traffic signals that responded in real time to pedestrian, cyclist, and vehicle flows, using embedded sensors and computer vision
• Heated sidewalks that automatically melted snow and ice based on weather sensor data
• Dynamic building facades with modular panels that adjusted to weather conditions and community needs
• Autonomous vehicle integration with dedicated lanes managed by real-time sensor networks
• Underground freight delivery systems optimized by logistics algorithms
• Continuous environmental monitoring of air quality, noise levels, humidity, and temperature at block-by-block resolution

Every one of these features depended on pervasive data collection. Sensors in sidewalks, lampposts, building facades, and public spaces would generate a continuous stream of information about how people moved through, used, and experienced the neighborhood. The fundamental question was immediate: Who would control all of this data?

The Proposed Data Trust

Sidewalk Labs recognized that data governance would be the decisive issue. In its Master Innovation and Development Plan (MIDP), released in June 2019, the company proposed a Civic Data Trust — an independent entity that would:

1. Classify all data collected in the Quayside area into categories: personal data, de-identified data, aggregate data, and "urban data" (a new category Sidewalk Labs invented for data about the built environment that was not directly personal but was generated in public spaces)
2. Require approval from the trust before any entity — including Sidewalk Labs — could collect, use, or share data within the project area
3. Operate independently of both Sidewalk Labs and Waterfront Toronto, with its own board of directors and legal authority
4. Apply a "Responsible Data Use Assessment" to every proposed data application, evaluating privacy risks, public benefit, and proportionality
5. Ensure data portability — no single company would have exclusive access to the data generated in the neighborhood

The proposal drew on the data trust concept described in Section 3.3.1 of this chapter: an independent trustee managing data on behalf of defined beneficiaries with fiduciary obligations. On paper, it was sophisticated, even groundbreaking.

The Backlash

The proposal immediately encountered fierce opposition from privacy advocates, academic researchers, community organizations, and former government officials. The objections clustered around several themes:

1. The fox guarding the henhouse. Critics argued that Sidewalk Labs — a Google subsidiary — was proposing a governance framework for data that Google itself would generate and benefit from. Ann Cavoukian, a globally recognized privacy expert who had been appointed as a privacy consultant to the project, resigned in October 2018 after learning that the data trust would not require all data to be de-identified at the point of collection. "I imagined us creating a Smart City of Privacy, as opposed to a Smart City of Surveillance," she wrote in her resignation letter.

2. The "urban data" category. Sidewalk Labs proposed a new data category — "urban data" — for information collected in public spaces. Critics argued this was a deliberate maneuver to create a classification that fell outside existing privacy law protections. If sensor data about pedestrian flows, vehicle movements, and ambient noise levels was classified as "urban data" rather than "personal data," it would be subject to the trust's rules but not to the full protections of Canadian privacy legislation (PIPEDA). Bianca Wylie, a civic technology advocate who became one of the project's most prominent critics, called it "a made-up legal category designed to let Alphabet do what it wants with data collected in public spaces."

3. Consent and the public realm. The Quayside neighborhood would be home to thousands of residents and visited by many more. Could residents meaningfully consent to pervasive sensor monitoring as a condition of living in their neighborhood? Could visitors? The project raised the question of whether data collection can ever be truly voluntary when it is embedded in the physical infrastructure of daily life — when the alternative to being monitored is not using a public sidewalk.

4. Power imbalance. Sidewalk Labs had committed hundreds of millions of dollars to the project. Waterfront Toronto had a small staff and limited technical capacity. Community groups had even fewer resources. Critics argued that regardless of how the trust was designed on paper, the structural power asymmetry between Alphabet and the community would ensure that Alphabet's interests prevailed.

5. Scope creep. The original agreement covered 12 acres. In its MIDP, Sidewalk Labs proposed expanding the project to 190 acres — a dramatic expansion that many saw as evidence that the company's ambitions far exceeded the original scope and that governance mechanisms would be perpetually outpaced by corporate growth.

Key Actors and Stakeholders

Sidewalk Labs / Alphabet: The project developer, with deep technical resources, significant financial investment, and a corporate parent whose core business model is data-driven advertising. Sidewalk Labs argued that the data trust would ensure community benefit and privacy protection. Critics noted that Alphabet's business model depends on data accumulation and that no governance mechanism proposed by a Google subsidiary could be trusted to genuinely constrain Google's interests.

Waterfront Toronto: The tri-government agency (federal, provincial, municipal) that managed the waterfront revitalization. Waterfront Toronto was the nominal partner and client, but its small staff and limited budget created a significant capacity asymmetry with Sidewalk Labs. Board members later acknowledged that the agency had been outmatched.

Toronto residents and community organizations: The people who would live with the consequences. Organizations like #BlockSidewalk and the Toronto Open Smart Cities Forum organized community resistance, testifying at public hearings and publishing analyses of the data governance proposal. Their concerns centered on consent, power, and the precedent the project would set for smart city data governance globally.

Ann Cavoukian: The former Information and Privacy Commissioner of Ontario and architect of the "Privacy by Design" framework. Her appointment gave the project credibility; her resignation gave the opposition legitimacy.

Bianca Wylie: A civic technology advocate and founder of Tech Reset Canada. Wylie became the most prominent public critic of the data trust proposal, arguing that data governance for public spaces should be determined through democratic processes, not designed by private companies.

Canadian policymakers: Federal and provincial legislators who held hearings on the project's data governance implications. The case contributed to broader Canadian debates about smart city regulation and amendments to PIPEDA.

Analysis Through Chapter Frameworks

The Stakeholder Map (Section 3.1.2)

Mapping the Sidewalk Labs case onto the stakeholder framework from Section 3.1.2:

The case demonstrates what happens when the algorithm builder's corporate parent is also the world's largest data company. The fiduciary obligation of a data trust depends on the trustee's genuine independence — an independence that critics argued was structurally impossible when the trust was designed by and accountable to an entity within Alphabet's corporate family.

Theories of Data Ownership Applied

• Data as property: Under this framework, the question is who "owns" the urban data. Sidewalk Labs invested in the sensors; residents invested in their neighborhood by living there. The property framework does not resolve this cleanly.
• Data as labor: Residents generate data through their daily movements. Under Lanier's framework, they should be compensated. But the Sidewalk Labs proposal did not include compensation — it proposed governance, not payment.
• Data as rights: The rights-based framework suggests residents have an inalienable right to control data about their lives. The right to informational self-determination cannot be contracted away, and pervasive sensor monitoring in public spaces may violate this right regardless of governance structures.
• Data as commons: The commons framework, drawn from Ostrom, suggests that urban data should be governed by the community that generates it. This was, in theory, what the data trust aimed to accomplish — but the community argued it was not genuinely in control.

The Data Trust Test

The Sidewalk Labs case became a test of whether data trusts, as described in Section 3.3.1, can work in practice. The concept — an independent trustee with fiduciary duties managing data for beneficiaries — is elegant. But the case revealed critical preconditions that the Sidewalk Labs proposal did not meet:

1. Genuine independence. A data trust must be independent of the entities whose data practices it governs. A trust designed by Sidewalk Labs to govern Sidewalk Labs data lacked structural independence.
2. Democratic legitimacy. The trust was proposed by a private company, not established through a democratic process. The community was invited to comment on a pre-designed structure, not to design the structure itself.
3. Technical capacity. The trust would need technical expertise to evaluate Sidewalk Labs' data practices, audit compliance, and assess re-identification risks. It was unclear how the trust would acquire and maintain this capacity.
4. Enforcement authority. A trust that lacks the legal power to compel compliance, impose penalties, or revoke data access is a trust in name only.

The Outcome

In May 2020, Sidewalk Labs announced it was canceling the Quayside project, citing the economic uncertainty created by the COVID-19 pandemic. Many observers argued that the pandemic was a convenient exit from a project that had become politically untenable due to unresolved data governance concerns.

The cancellation did not end the conversation. The Sidewalk Labs case became the most widely cited example in debates about smart city data governance worldwide. It demonstrated that:

• Data trust design matters less than data trust legitimacy
• Communities will resist data governance frameworks they did not help create
• The identity of the trustee is as important as the structure of the trust
• Data governance for public spaces raises fundamentally different questions than data governance for voluntary digital services

As Dr. Adeyemi might observe: the Sidewalk Labs data trust was a technically sophisticated answer to a question the community had not been allowed to ask.

Discussion Questions

1. The independence problem. Could a data trust proposed by Sidewalk Labs ever have been genuinely independent of Alphabet's interests? What structural safeguards — if any — could have made it credible? Is there a model from another domain (financial regulation, judicial appointments, public broadcasting governance) that could inform data trust design?

2. The "urban data" question. Sidewalk Labs proposed a new data category — "urban data" — for sensor data collected in public spaces. Is this a legitimate and useful classification, or was it (as critics alleged) a deliberate attempt to circumvent privacy protections? What governance rules should apply to data collected by sensors in public spaces about people who cannot practically opt out?

3. Consent in the built environment. Can people meaningfully consent to data collection embedded in the infrastructure of the neighborhood where they live? How is this different from consenting to terms of service for an app you choose to download? What governance mechanisms could protect residents who did not choose to live in a "smart" neighborhood?

4. The precedent question. The Sidewalk Labs case was watched globally as a test of smart city data governance. If the project had succeeded with its proposed data trust, what precedent would that have set? If the precedent was problematic, was cancellation the right outcome — or should the project have been reformed rather than abandoned?

Your Turn: Mini-Project

Option A: Data Trust Redesign. Assume that a different developer — not a subsidiary of a data company — proposed a smart neighborhood for the Quayside site. Design a data trust governance framework that addresses the concerns raised by critics of the Sidewalk Labs proposal. Your design should specify: (1) who serves as trustee and how they are selected, (2) what data is collected and how it is classified, (3) what approval process governs data use, (4) how residents participate in governance, and (5) what enforcement mechanisms exist. Present your design in a two-page proposal.

Option B: Comparative Smart City Analysis. Research one other smart city data governance initiative (e.g., Barcelona's "data sovereignty" approach, Singapore's Smart Nation program, Amsterdam's Responsible Sensing Lab, or Helsinki's urban data platform). Compare its data governance model to the Sidewalk Labs proposal. Where does the alternative model address the problems that derailed Quayside? Where does it face similar challenges? Write a two-page comparative analysis.

Option C: Community Voice Simulation. Imagine you are a resident of the proposed Quayside neighborhood. Write a two-page public comment on the data trust proposal, drawing on the frameworks from Chapter 3. Your comment should: (1) identify which aspects of the proposal you support and why, (2) identify which aspects concern you and why, (3) propose specific modifications using at least two of the four ownership theories, and (4) articulate what governance structure you would need to see before consenting to live in the neighborhood.

References

• Sidewalk Labs. Master Innovation and Development Plan. Toronto: Sidewalk Labs, 2019. Available at: sidewalktoronto.ca.

• Sauter, Molly. "Google's Guinea-Pig City." The Atlantic, February 2019.

• Wylie, Bianca. "Searching for the Smart City's Democratic Future." Centre for International Governance Innovation (CIGI), 2018.

• Cavoukian, Ann. "Letter of Resignation from Sidewalk Labs Advisory Panel." October 2018.

• Balsillie, Jim. "Testimony before the Standing Committee on Access to Information, Privacy and Ethics." Canadian House of Commons, 2019.

• Scassa, Teresa. "Designing Data Governance for Data Sharing: Lessons from Sidewalk Toronto." Technology and Regulation (2020): 44-56.

• Goodman, Ellen P., and Julia Powles. "Urbanism Under Google: Lessons from Sidewalk Toronto." Fordham Law Review 88, no. 2 (2019): 457-498.

• Sidewalk Labs. "Why We're No Longer Pursuing the Quayside Project." Medium, May 7, 2020.

• Wylie, Bianca, and Sean McDonald. "What Is a Data Trust?" Centre for International Governance Innovation (CIGI), October 2018.

• Open Data Institute. "Data Trusts: Lessons from Three Pilots." London: ODI, 2019.