We live in an era of big data, which is increasingly used by powerful actors to advance a range of financial, political and military objectives. In just the past few years, we have witnessed how multinational corporations and governments have abused such data: from the Cambridge Analytica scandal, revealing how personal data were mined from Facebook pages without users’ knowledge or consent to promote both the Brexit campaign and Donald Trump’s US presidential bid, to signature drone strikes that target people based on algorithmic calculations of digital signals.
For those interested in social justice, the transformation of private aspects of our lives into data, which are then used for profit or for violent ends, presents immense new challenges and urgent questions. How can we regulate the way data are gathered, produced and used? And how can we mobilise data to improve the lives of the many rather than serve the interests of the few?
In Data Feminism, Catherine D’Ignazio and Lauren Klein begin addressing these pressing questions.
They start by demonstrating how data science has historically been informed by endemic sexism and racism, and showing that no data – even raw data – are ever neutral or objective.
To drive this point home, the authors provide numerous examples, including the well-known case involving Joy Buolamwini, a researcher at the MIT Media Lab, who discovered that most facial recognition software does not recognise darker-skinned or female faces. The book also details less-known examples, such as a risk assessment algorithm developed by the company Northpointe. This algorithm produces “scores” meant to evaluate the risk that criminal defendants pose – scores that influence whether a person is sent to jail or set free. Yet when investigative reporters examined how the algorithm operates, they discovered that black defendants were (mis)labelled “high risk” far more often than white defendants.
These and other cases highlight the insidious ways in which racial and gender prejudice are baked into artificial intelligence and algorithmic technology.
Rather than dismiss data as inherently biased, however, D’Ignazio and Klein advocate for what they call “data feminism”. They define this as “a way of thinking about data, both their uses and their limits, that is informed by direct experience, by commitment to action, and by intersectional feminism thought”. Doing data feminism means basing one’s approach on seven principles: examine power; challenge power; elevate emotion and embodiment; rethink binaries and hierarchies; embrace pluralism; consider context; and make labour visible.
This feminist approach understands that data are shaped by and through particular power relations, while its objective is to facilitate the design of data systems that are inclusive, responsive to community needs and geared towards improving people’s lives. Data feminism, in short, reclaims datasets and systems as social justice resources for communities.
Many of the arguments in Data Feminism are not particularly new. After all, feminist critiques have long challenged scientific claims to objectivity and neutrality. However, the authors’ demystification of data science and advocacy for data feminism are extremely timely. The book also serves as an important introduction to intersectional feminist practice by providing inspiring examples of marginalised women and communities taking power back by collecting and wielding “counter-data” to challenge the status quo.
Catherine Rottenberg is associate professor in American and Canadian studies at the University of Nottingham and the author of The Rise of Neoliberal Feminism (2018).
Data Feminism
By Catherine D’Ignazio and Lauren F. Klein
MIT Press, 328pp, £25.00
ISBN 9780262044004
Published 30 March 2020
