Architects and designers ignore data at their peril. It is as important as any other building component, writes Sarah T Gold
The built environment comprises materials, facades, plans and people. Increasingly, it also comprises data. Retail shops track shoppers’ movement and even what draws their attention. WeWork offices are full of sensors that track and improve how spaces are used. Transport for London provides free WiFi on the Underground and collects the data you use. Amazon’s smart doorbells share video with law enforcement. And the Information Commissioner’s Office is currently investigating a property developer’s use of face recognition cameras at King’s Cross station in London.
Do you know what data your Nest thermostat collects? Or that smart speaker in your friend’s home? Or the smart entry system in that office you’re visiting? It’s hard to tell which video, audio and data are being collected. And it is even harder to tell where that data is being sent, who will use it and for what purpose.
When computers and buildings have sensors and internet connections – which in cities is increasingly a great deal of the time – there is always uncertainty. This uncertainty exists for several reasons. Consumer technology companies often have commercial incentives to collect data. Online services change over time and therefore so does the way they use data. Moreover, software supply chains are complex and opaque. Yet we still ask individuals to make decisions about data permission when they can only see the software supply chain from the far end, and dimly.
So far, I’m describing well-known problems with the internet and software. But this matters to the built environment too because ‘software is eating the world’, as one Silicon Valley venture capitalist, Marc Andreesseen, approvingly put it in a Wall Street Journal article in 2011. Software is now finding its way into urban planning, civil engineering, architecture and smart building management.
Permission to access the internet is generally constructed as a one-to-one relationship between the purchaser of a device and its manufacturer, but in most places where people live and work, devices are shared. This poses risks when one person sets the terms and permissions for everyone around them. Permission is often requested once only, or in an unclear way. Consent is often reduced to a ‘click through’ barrier, rather than something understandable and considered. And even though services typically promise to anonymise and keep safe personal data, that data may still be aggregated in a way to re-identify individuals, as celebrities using New York taxis found in 2014.
Imagine what buildings and cities would be like if they worked like websites. You might have to click through a lot of terms to get through a front door. The public might be monitored and measured to increase their happiness or just their dwell times. Buildings might improve with time or be constantly reconfigured. Neighbourhoods might be instrumented to improve their efficiency, security or profitability. Buildings might be more needy, a mess of things trying to get your attention. Your TV would adapt to your watching habits but your fridge might be squabbling with your door lock..
That’s to say: data is changeable, uncertain, opaque. You don’t know where data is kept or how, and data changes over time. Data flows through a complex web of interest and power, permission and trust, use and commerce usage. Data can be recombined infinitely to make more data. Because of this complexity and malleability, data has many many stakeholders. Building owners have the right to install CCTV, sensors; but people also have rights under data protection legislation, and there’s an overlap and tension between these rights. Therefore, as soon as buildings have people in them, the webs of permission and trust are already complicated. The built environment must acknowledge that complexity.
I believe that architects and engineers must think of data as a material. They must start with visibility: by disclosing data collection in a way that makes it clear and understandable. The built environment should reveal the ‘seams’ of its data. Just as some buildings have information displays that indicate energy performance, a building might show how it collects and uses data.
There should be open registers of sensors in the public and semi-public realm. What if a building let you query it - or the organisations it houses - for data held about you? Could it do that while still respecting your privacy? People should have agency over their permission choices, and there should be avenues for users to hold whoever is collecting the data to account.
We design buildings for many modes of human interaction - meeting rooms, shared spaces, private areas, serendipitous corridors. Buildings could also be designed for different modes of data, with zones that explicitly offer a luminous bath of real-time analytics and other that are cones of data privacy.
Underlying this approach are some big questions. What should people’s digital rights in the built environment be? Who should manage those rights and how? We can begin to answer those questions if we treat data as a material for building with.