When Biometrics Fail: Gender, Race, and the Technology of Identity

In the course of the past decade we have all, on one occasion or another, encountered situations where we were asked to provide a fingerprint, have our picture taken, look into an iris scanner, or have some other bodily feature digitally registered, stored and used in systems whose workings remained highly opaque but were said to be for our own good. Our security in particular is said to increase tremendously when we allow our bodies to be used for identification purposes. Biometric identification technologies - biometrics for short - have grown from being a solution looking for a problem into a multibillion-pound industry that claims to provide answers to a range of societal insecurities and threats. From terror and crime to social security fraud and illegal migration, these technologies are advocated and presented to us as necessary to fight great dangers. Thus we are encouraged to accept that billions in public funds (and corresponding profits for their producers) should be spent on systems that are promised to make us more "secure" (in any of the many meanings this term covers today).

Shoshana Amielle Magnet's When Biometrics Fail joins the voices questioning these claims and their underlying assumptions and drivers. Technology scholars working on biometrics have argued that while such measures do not deliver on their advocates' promises - they fail systematically to identify correctly - they do something else effectively: they reinforce ethnic, racial, gender, age and other inequalities.

Moreover, these two aspects are closely related: it is partly through the specific way these technologies are failing that inequalities, exclusions and injustices occur. Biometrics systematically work better for some segments of the population than others. When entitlements, services or, for example, passage through security checkpoints at airports are made contingent upon successful biometric identification, as they are to an increasing extent, this means that segments of the population are systematically disadvantaged. The widespread application of biometrics in areas with high stakes, such as crime prevention, social security, border control and asylum and migration management, makes this a non-trivial matter.

Taking her cue from science and technology studies' methods and theories, where definitions of "success" in connection with technological developments are long-standing topics of interest (what does it mean that a technology is claimed to be successful?; whose definition of "success" is this, and whose perspectives does that exclude?), Magnet focuses on its counterpart, the issue of technological failure. In view of the fact that their often substandard performance seldom seems to play a role in government decisions on whether to opt for the large-scale implementation of biometrics (eg, the US-Visitor and Immigrant Status Indicator Technology system, or Europe's use of biometric passports), exposing whole populations to the consequences of their failure, these are timely questions.

Unfortunately, Magnet's answers are somewhat disappointing. Saying that biometrics will not "solve the complex problems of poverty, homophobia, ableism, sexism, and racism", or that they fail to "represent bodily complexity", will probably not surprise too many readers, because I doubt anyone ever believed they could; arguing that we should stop investing "billions of dollars" on technological solutions to social problems - even if for many of the large-scale systems that approach would probably be well advised - wrongly suggests that some "real nature" of the problem makes biometrics the wrong solution.

Having said that, for those interested in the technology, its social implications and use in a number of US contexts, the details of Magnet's investigations are very much worth reading.