What Is a Data Trust and Why Are We Even Talking About It? Sidewalk Labs’ Magic Tricks

Posted January 14, 2019
By Mariana Valverde

Sidewalk Labs, a unit of Google’s parent company, with a big assist from Waterfront Toronto, started their career in magic over a year ago. They directed our collective attention to artsy pictures of imaginary buildings and gadgets – and away from anything financial. And anything legal. 

Instead we get terms that sound legal but are very murky – like ‘data trust’, the notion proposed by Sidewalk Labs to counter growing concerns about privacy and data commercialization. I’m not sure why the word ‘trust’ would be used to convey public mindedness, since in Canada the main users of trusts are rich people leaving money to offspring while minimizing taxes. Maybe it’s just the word itself, however: wouldn’t a trust just be trustworthy?

Perhaps Sidewalk hoped that their proposal would bask in the virtuous halo of the marginal legal category of charitable trusts. But charitable trusts are very regulated, and have to benefit a defined enterprise that qualifies as ‘charitable’; and whatever happens in the Quayside district, it is unlikely to qualify as a charity. 

In general, a trust consists of assets, generally in private ownership, administered by a person or institution taking on the trustee role. Trustees administer the asset not for personal gain but for the benefit of an identifiable group, the beneficiaries: they owe beneficiaries a ‘fiduciary duty.’ So a trust can benefit a group – a group that by definition does not own the asset, it’s important to note -- but only if the asset’s owner decides to do so. And a trust must have identifiable beneficiaries. 

That’s a problem, since nobody knows who could be the beneficiary of the proposed Sidewalk data trust: Waterfront Toronto itself? The City of Toronto? Waterfront residents? City residents? What about visitors whose data are harvested while sunbathing at Sugar Beach? 

Ordinarily a trust’s beneficiaries are either named individuals – the five grandchildren of Mr. Rich, say—or an identifiable group (such as patients receiving services at a particular place).

Canadian law can’t tell us who would be the appropriate beneficiary for a data trust; but it does tell us that the owners of the assets are the ones who decide who the trustees will be. And Canadian law tells us something else that Sidewalk is hiding: that the general public is the one group that cannot be a beneficiary of any trust. 

Transit agencies and public hospitals serve the public at large. But these bodies are not trusts; they are public corporations, with publicly owned assets managed by elected or appointed authorities accountable to governments and ultimately to the voting public. Waterfront Toronto was created as just such a public corporation, of course; but it has been acting as Sidewalk’s enabler rather than as champion of the public interest – as shown by their continuing silence on the question of intellectual property.

Just this week, the Toronto Board of Trade (BOT) has started to ask the questions Waterfront Toronto should have publicly asked, in their January 2019 report Biblio Tech. The report argues that the Toronto Public Library, which has considerable expertise in information technology, should be the public actor that engages with Sidewalk in regard to data and digital issues. Replacing Waterfront Toronto by the Toronto Public Library in dealings with Sidewalk on digital matters would certainly be an improvement –although the library and/or the city is unlikely to have the specialized legal advice needed to negotiate. 

But the Board of Trade goes further: they are unimpressed by the data trust notion, and suggest a data “repository” instead. A repository might be a more democratic and transparent mechanism for data governance than a trust; more qualified folks should weigh in on that. What I can do is help to put a nail in the coffin of the Sidewalk data trust from my perspective as a legal studies scholar.

The first thing to note is that it was Sidewalk that got us to talk about data trusts in the first place. Why was that? Well, if citizens are told, ‘look at the data trust, over here’, we will not see what the magician’s other hand is doing – which is getting rich by putting data that we generate through their software to produce wealth, American wealth. 

Sidewalk has talked about some principles for a data trust. But de-identifying data doesn’t make it public, it just steers commercialization into certain channels; and making privately owned data sets available doesn’t make them publicly owned. If letting citizens and governments have some access to privately owned data sets is what the data trust amounts to, the questions that arise from my reading of the Board of Trade report are: Why only data? And why a trust, which can only be set up by whoever owns the assets in question?

If privately owned data sets are opened to public access (which isn’t as democratic as it sounds, since only tech companies or governments have the capacity to make use of large data sets), Sidewalk and its mysterious subsidiaries can just quietly hoard the really valuable algorithms and software inventions --while citizens discuss privacy settings, which is what Sidewalk wants us to focus on. Questioning Waterfront Toronto’s subservient acceptance of ‘data governance’ as the only issue, the Board of Trade report acknowledges that intellectual property assets will be the key wealth generated at Quayside. Without arguing for full public ownership -- they are the Board of Trade and not the NDP, after all—the report’s authors make the common sense observation that the public ought to benefit financially from any digital innovations to which Torontonians contributed, knowingly or not. 

The Biblio Tech report brackets many important questions, such as land ownership, land value, construction financing, and, to go back to the beginning of our drama, the questionable decision by Waterfront Toronto to issue a Request for Proposals when they didn’t know what they wanted to procure. As citizens discuss the Biblio Tech report, finally realizing that there are options beyond a ‘data trust’, Sidewalk-led or not, let’s try to avoid the common mistake of staring at the magician’s proffered hand so earnestly that we let them get away with more tricks.

 

Thanks to Lisa Austin, Alex Flynn, Neve Peric and Bianca Wylie for many fruitful discussions, though none are responsible for the views expressed here.