‘People still think of sensors as devices that will steal their data’

By 2025 there will probably be a trillion sensors on the planet. It’s not an exact figure, but for Chris Slater it hardly matters: they’re going to outnumber humans by orders of magnitude. What matters more to the author of ‘Sensing Machines’ is that the public – who are becoming more reliant on the interconnected world of sensing devices – don’t really understand them.

“People still think of sensors as devices that will steal their data. But they’re just transducers that take a signal out of the world and change it into another form of energy so that a computer can read it.” What people are thinking of, he says, is the machines connected to the sensors; they’re thinking about the corporations that harvest data to create wealth.

For Salter, sensors are the bridge that provides the interface between the digital and human worlds. “Think of it like a human body,” he says. “The functions of the sensors – the eyes, ears, skin and so on – make no sense unless they’re connected to the nervous system and the brain.”

Today’s electronic sensors are “deeply entangled, invisible and ubiquitous, and they are in devices that we directly interact with. Nobody thinks of game controllers as having sensors in them. But what is it that makes you feel you’re part of a game? It’s the sensors. Back in 2000, after a lot of experimentation, Nintendo put accelerometers in the controller. The minute that happens, you have a different relationship between the digital world and your body. Suddenly your body becomes not just an extension of the technology but the technology itself.”

‘Are these new technologies being designed to reflect how humans act?’

What we have here, says Salter, is a case where a sensor is measuring something “mathematically abstract” and interacting with us. “That’s what changes the orientation of the displays on our phone screens. It’s what stops your hard drive if you drop your laptop.” All this physics brings machines “closer to how we are in the physical world”.

Electronic sensors came into their own in the 21st century, says Salter, when “computer science and sensing technology left the laboratory and became part of everyday life”. You can go back to the 1950s, he says – ‘Sensing Machines’ does, and further to the dawn of cybernetics and artificial intelligence – “but at that time this was only being worked on by a tiny handful of scientists. Of course, people like Claude Shannon, who was thinking about the mathematical theory of communication at Bell Labs, knew that these things would have a big impact down the line. They all knew that digitalisation was around the corner. But there were only a few people involved. Today, it’s everywhere. Sensors are in everything we own and everything we encounter. You might say that engineers won.”

This victory, says Salter, is accompanied by an “interesting problem”, which is the central point of ‘Sensing Machines’. “As they became ubiquitous, people started to realise that these technologies were not neutral. Engineers engaged in machine learning today realise that what they’re working on has an impact on the world – their mathematics is actually making a material transformation in the world. It’s not some abstract logic model that sits inside an academic paper. For example, today there is a big debate about fairness and unconscious bias in AI, machine learning and all sorts of sensing technologies. And this is huge because suddenly engineering has to be critical of itself – something that engineers are not taught.”

This shift from “solving problems better to having better critical reflection” is something that needs to be examined, says Salter. Engineering modern sensing systems is creating problems too, especially when it comes to knowing “what model of human being is being considered when it comes to the design. We need to ask if we are looking at humans simply as robot-like cognisant entities. Or are these new technologies being designed to reflect how humans act and behave in the world?”

The opening of ‘Sensing Machines’ is a film scenario, tracking a human in a socio-economically advantaged society. Our human is immersed in a world of smart speakers, newsfeeds, CleverZones, financial data collection, wearable electronics, brainwave monitoring, self-driving cars, AI climate control, bio-medical data capture, web-based security, and of course an intelligent coffee machine that delivers a brew of the perfect temperature and strength. The scene follows the human to work, to a gallery, to an exercise regime, through to finally setting the parameters on a wrist sleep tracker.

The author’s lens zooms out to encompass several global challenges being met with sensor resistance: pandemics, climate change, the built environment. It’s all good in the digital world. Apart from the fact that it might not be, because of the subconscious bias mentioned earlier: “Supposedly ‘objective’ sensors perpetuate long-​ingrained human biases. Soap dispensers using infrared sensing to detect motion favour lighter skin. Facial-recognition software mainly identifies white faces. Alexa and Google Home’s microphones discriminate against Indian and Black English.” There’s also the spectre of how we use the data to decide “who/what would be classified as ‘normal’ and who/what would deviate from the norm”.

This last point brings Salter onto the use of big data to target individuals by analysing mass patterns “not only to count and quantify, but to predict”.

There’s already an Orwellian lexicon evolving to describe what Salter calls ‘prediction culture’: surveillance capital, dataveillance, knowing capitalism, datafication and extractive capitalism. “These predatory techniques depend on the vast digital information infrastructures possessed mainly by the behemoth FAANG corporations (Facebook, Amazon, Apple, Netflix, and Google): mass deployments of sensors, artificial intelligence, server farms, data centres, marketers and advertising agencies to capture and shape our behaviour.”

One of the key strengths of ‘Sensing Machines’ is its editorial balance. On the one hand, Salter describes a scary landscape of corporate giants using sensors to harvest our data in ways that consumers simply don’t understand. But on the other, there is a whole new world of perception and sensation just waiting to be explored.

‘Sensing Machines: How Sensors Shape Our Everyday Life’ by Chris Salter, is from The MIT Press, £25