The government estimates that location data could add £11 billion to the UK’s digital economy each year, but the insights it provides can have alarming consequences. Ukrainians got a taste of this in 2014. Demonstrators against then-President Viktor Yanukovych claimed that the government had tracked location data in mobile phones to send an ominous message to people in the vicinity
The government estimates that location data could add £11 billion to the UK’s digital economy each year, but the insights it provides can have alarming consequences.
Ukrainians got a taste of this in 2014. Demonstrators against then-President Viktor Yanukovych claimed that the government had tracked location data in mobile phones to send an ominous message to people in the vicinity of a clash between police and protesters: “Dear subscriber, you are registered as a participant in a mass disturbance.”
The risks are typically less dramatic but insidious nonetheless. A 2018 University of Oxford study estimated that 90 percent of the apps we use have third-party trackers embedded into their source code that collect our location and other data to infer our habits and beliefs. This information creates detailed profiles that can be used for micro-targeted political advertising and credit scoring.
“Increasingly I find people saying I don’t want to turn on location services because I don’t know what it means,” says Alex Wrottesley, head of Geovation, Ordnance Survey’s incubator for geolocation startups. “And that’s a problem if you want to build businesses that leverage those datasets.”
In response, Geovation has launched the Benchmark Initiative, a year-long entrepreneur-in-residence programme and lecture series that aims to promote public awareness and ethical applications of location data.
Among the speakers at the second Benchmark lecture was Dr Anahid Basiri, a spatial data scientist at UCL who has been researching ways of adapting location-based services to the needs of different users. One of her studies compared the benefits of 3D and 2D maps. Basiri discovered that women preferred to use 3D maps as they could use buildings and landmarks to navigate their paths, whereas men favoured 2D versions.
Basiri traces the gender difference to the origins of cartography, when prehistoric men would rely on mental maps to plot their hunting routes.
Millenia later, the trend still lingers in crowdsourced location services such as OpenStreetMap, which Basiri estimates has a contributor network that is around 95 percent male. As a result, the maps they generate visibly lack representation of the experiences of women.
“For example, there is only one type of childcare service in the whole map,” Basiri tells Techworld. “Nobody distinguishes between a nursery, primary school or any other childcare service, because traditionally it is the female’s responsibility to take care of the kids. But there is a very detailed source of information about nightclubs and strip clubs.”
Similarly, the lack of elderly, ethnic minority and disabled contributors means that the locations they are more likely to search for in maps are often absent.
Basiri and her colleagues at UCL have tried to help fill these gaps in maps by creating Colouring London, a web-based platform that enables citizens to contribute information about buildings in the city by colouring them in on an online map. The colour that they choose visualises the building age, use and size, and designation status in a simple interface that aims to democratise cartography.
Another of Basiri’s project explores whether consumers would compromise the accuracy of personalise advertising by providing less location data. The results revealed a temporal pattern to user preferences. When users were near their home address, they were reluctant to share their location. But when they were close to a shopping centre, they were keen to provide it in order to receive the best deals available.
Basiri acknowledges that the impact of her research into user representation and preferences will be limited without support from developers. She urges them to consider the worst-case scenarios of any services they design, and the long-term effects that their decisions could have on representation and privacy.
“Sometimes the changes are quite easy to make,” she says. “It’s not a huge amount of effort that you need to just adapt it to a certain group of people. It’s just like language settings; it’s not that difficult. And all you need to have is just some APIs to make it language adaptive, or different screens for the elderly that are usually a bigger size.”
Mapping the future
Back at Geovation, Wrottesley is focused on educating users on what exactly they are consenting to when they allow a service to use their location data. He believes that this is best achieved through tools that show the consequences of different scenarios in a manner similar to how the MIT Media Lab’s Moral Machine tested the decisions they wanted self-driving cars to take when destined to crash.
Wrottesley believes that companies using location data are starting to take note of the concerns, but admits that their commercial priorities are limiting their actions.
“There needs to be some joint understanding between competitors as to how much is required,” he says. “It’s hard for anyone to take the first leap on this because there might be a commercial impact, and people obviously are cautious of that.”
Basiri has seen positive steps taken by services that involve vulnerable groups, such as those that track the movements of children to inform parents of their whereabouts, but is still waiting for a tech giant to provide a positive example of ethical location data use: “When it comes to the biggest money generators? To be honest, I look forward to seeing one of them.”