Culture Eats Technology for Breakfast
"Will we live in a beautiful Utopia or a dystopian Big Brother society?" Journalists, politicians, and even random acquaintances ask this question when they learn about my work with Big Data and cities. In fact, most researchers and practitioners working in the Big Data space have shared this experience.
Hollywood has shaped our image of a data-rich world, where all we need to do is: “Ask a question. Hit a button. Technology will magically solve your problem.” Indeed, our planet has become blanketed with sensors, apps, and new models to monitor, understand, and guide human behavior and nature’s phenomena. Three years ago, there was one closed-circuit television (CCTV) camera for every 11 people in the UK. The following year there were more SIM cards available globally than people to use them (a quarter of them enabled machines to talk to other machines). Today, more than 1,300 operational satellites orbit the earth, while 1.5 billion Facebook users and 540 million WeChat subscribers share their lives on social media.
For anyone with a dystopian outlook, this might seem cause for alarm. Especially because, in most movies, robot overlords finally dissolve our humanity and obliterate the human race itself – a narrative that elicits an exasperated gasp from researchers across all disciplines.
Scientists and technologists recognize the challenges of data collection, data analysis, and building control systems, even for simple real-world applications such as traffic light networks. At the same time, artists, philosophers, and social scientists work to understand the boundlessness of human creativity and curiosity, and the many perspectives and lenses through which diverse individuals and societies perceive the kaleidoscope of human life.
Despite this rich personal and cultural diversity, most experts and the general public continue to expect that technologies will work universally, with Big Data models describing general laws, and systems able to be scaled globally. However, my daily work has shown that those assumptions fail hard and often. I attribute this to a lack of conversation and collaboration between the engineers and the artists, between scholars in science and those in the humanities.
There has been widespread failure to understand and react to cultural specifics and a tendency to underestimate the degree to which the operation of sensors and software is based on factors rooted in the culture and expectations of their inventors. These culture-based factors may ultimately have severe impacts on people’s daily lives.
As a scientist at MIT Media Lab and an urban planner in Asia, I try to understand human mobility patterns in order to (re)build cities with better urban infrastructures for all citizens, particularly the poor. This work has revealed that several of the Western-developed methods and tools relying on Big Data cannot be directly transferred and used to understand movement in Asian countries. For example, large datasets collected from mobile phones are currently explored to understand daily mobility patterns in cities. But the methods developed to evaluate these data were based on assumptions intrinsic to the (Western) societies where they originated.
Among the Western cultural expectations are: a phone always belongs to the same individual, who keeps it with him/her at all times; the telephone is not shared with other members of the family; a phone is not frequently bartered to get money for more urgent needs, etc. It is also assumed that people always sleep in the same spot and the phone rests there for the night, identifying a stable “home” location. Yet, many of these assumptions do not hold true for poorer, Asian societies, so analytics should not be applied identically.
Another reason why Western analytics often fail when applied to Asian societies is that most mobile phone contracts in Asia are pre-paid. Post-paid contracts are a “luxury” for the rich. In the People’s Republic of China and Southeast Asia, for example, 85% of all connections were prepaid in 2013, as were 95% of all connections in India, Pakistan, and Afghanistan. Therefore most mobile phone operators in poor cultures do not need to store data from each call for billing information and thus save on storage costs and billing software.
Thus, if the data available from service providers (this is still Big Data from millions of calls) are analyzed to understand the transport behavior of an urban population, it only captures data from the rich who can afford bank accounts, and post-paid contracts. Clearly these data do not reflect mobility flows of the poor or even of typical citizens. Governments that base new public transport lines or other urban services on such Big Data analytics will fail spectacularly – mostly because we, as researchers, failed to collaborate across disciplines when developing the methods and models for interpreting Big Data.
Long before Hollywood’s “Artificial Intelligence as humanity’s doom” narratives come true, we must learn to limit the number of culturally incompatible projects based on Big Data – especially when implementing projects in the most vulnerable societies for the most vulnerable people.
Big Data and new methods, such as network science, can be employed to better understand the world and its people(s). But first we must recognize that technology is not culturally agnostic. At this year’s European Forum Alpbach in Austria, we started this multi-faceted conversation across our different disciplines with students, researchers, politicians, and journalists from three continents in a week long seminar on "Understanding Big Data." I look forward to continuing those conversations at the ARIT in Toronto.
This article builds on the chapter “Big Data for Urban Change – Debunking the Myth & a Way Forward” in the forthcoming book: Urban Change, Anton Falkeis, ed., De Gruyter, 2016.