Who Controls the City When Software Is Privately Owned?

Who controls the city when software is privately owned is the question left by the failure of Sidewalk Labs in Toronto: in 2017 Canada's prime minister came down to the waterfront to announce that an urban subsidiary of Google would design an entire neighbourhood, built, in its own words, from the internet up, with heated pavements, driverless taxis, traffic lights that learn from traffic and, beneath it all, thousands of sensors recording data on every person who crossed the street. That neighbourhood was never built, and its failure matters more than many successes, because it exposes the underlying issue: when the programs that run a city belong to a company, does the council that signs the contract govern, or the vendor that writes the code.

The mechanism: a purchase of services that is a transfer of power

The smart city model almost always works the same way. A council buys from a vendor a platform promising efficiency in traffic, irrigation, lighting or security; the company installs the sensors, operates the servers and designs the algorithms that turn data into decisions. On paper it is a purchase of services; in practice it is a transfer of power. Urban data start circulating through private infrastructure, the criteria that prioritise one avenue or neighbourhood are protected as trade secrets and changing vendor becomes so costly and complex that the city is locked in. Specialists call it vendor lock-in: the public client can no longer leave.

Algorithmic opacity: decisions nobody voted for inside a black box

The effects appear once there is no way back. Decisions once debated in a council chamber, which street gets patrolled more, which light turns green first, which neighbourhood gets priority maintenance, are executed inside a black box nobody voted for and someone programmed in a distant office, and when a councillor asks to audit the system the usual answer is that the code is protected by intellectual property. This is algorithmic opacity, which turns urban politics into an outsourced technical service, and the big contracts also concentrate the market in a handful of global corporations. The city, in theory everyone's, starts running on rules almost nobody can read.

Toronto: Sidewalk Labs, Balsillie and the question of data

Toronto is the most revealing case. Sidewalk Labs was not content to install technology: its founding documents proposed charging fees, running public transport and trialling its own forms of local policing and justice, and the Canadian tech entrepreneur Jim Balsillie called the project a colonising experiment in surveillance capitalism. Organised citizens responded with hearings, protests and formal petitions asking who would own the data and to whom the system would answer; the company never gave a convincing answer and abandoned the project in 2020 without laying a brick. Toronto showed that the model can be stopped and how exhausting stopping it is.

Songdo, Zuboff, Morozov and the municipal alternatives

The second case is in South Korea. Songdo was raised from scratch as the perfect smart city, wired, sensored and exhibited at trade fairs as the definitive urban future, and many researchers now describe it as a catalogue city, technically impeccable and socially lukewarm, working above all as a showcase for its own vendors. The indices that rank smart cities close the loop: they reward those that buy more technology and press the rest to imitate them, the index acts as catalogue, the vendor supplies the product and the subcontracted city is the result. Shoshana Zuboff reads it as surveillance capitalism carried into the street, extracting human behaviour and turning it into business, and Evgeny Morozov adds solutionism, the belief that every urban problem is solved by buying technology.

It can be read as a mutation of neoliberal urbanism: once water or bus companies were privatised, now the capacity to decide is outsourced. Alternatives exist and are municipal. Barcelona has included data sovereignty and free software clauses in its technology contracts since 2016; Amsterdam and Helsinki publish registers of the algorithms they use and for what; several cities require that data generated in public space be public and that any system be auditable and replaceable. Toronto's lesson is that the question of who owns the data decides who governs, and that a council that cannot read its city's code has stopped governing it, even if it keeps signing the contracts.