The power to make persons
Who gets to say who is?
Egged on in part by the widespread pushback against fake news, social media giants have become more proactive about “fact-checking” content.
In practice, “fact-checking” is a palatable way to describe the vast networks of AI and human labor that flag content as real or fake, true or false.With these networks, companies like Facebook get to decide what is true, what is authentic.
But what about the power to decide who is authentic?
Who decides who is?
States have a monopoly on “legitimate” identity. Records like birth certificates, passports, and visas (cl)aim to provide a canonical, authoritative, or ultimately trusted form of personhood. Predictably, in cases of migration and displacement, states strategically grant or deny those documents to mediate access to employment, education, mobilities—the interlocking networks of opportunity that Foucaude and Healy call “life chances.”
A similar dynamic plays out in digital identity systems.As with Instagram’s “fact-checking” algorithms, these systems often judge identity through probabilistic models. And these probabilities, in their particulars, are political. Jonas & Burrell’s (2019) work show systems designed to detect “weird” behavior in online platforms effectively flagging non-Western behavior. (Try logging into your bank account from Ghana or Cameroon).
And these probabilities can kill. In her essay A Sea of Data, Hito Steyrel tells the story of Amani al-Nasara, a woman permanently blinded by an aerial drone strike in Gaza.Of the incident, Steyrel asks:
What kind of “signal” was extracted from what kind of “noise” to suggest that al-Nasara was a legitimate target? […] Or to put it differently: who is “signal,” and who [is] disposable as “noise”?
It should be no surprise that similar algorithms are busy making judgments about the state-issued markers of personhood I mentioned above. IBM infamously applied its Watson AI system to label ‘terrorists’ among refugees.Meanwhile, India’s national identification project uses biometrics to delegate authentic identity—imperfectly, and with particular political visions in mind.
While digital accounts and legal documentation in a country do not make for a one-to-one comparison, both mediate the “authenticity” of one’s personhood and, in doing so, control access to services, opportunities, and mobility.They interact with, inform one another in structure and kind. And for both, the central debate to my mind is: who has the power to grant authenticity? Whose notions of authenticity get to matter?
The power to make persons
When people talk about the power of tech companies and the ways tech companies compete with states for power, they often gloss over tech companies’ ability to create and manage systems of identity. The identities tech companies give us matter. They shape life chances. What does it mean that Google controls the “ground truth” for my primary identity on the Internet—my email address?What does it mean when a Facebook account attests to someone’s authenticity on a dating or employment site?
Making identity systems (and deciding which identities are “real”) is a foundational vector of state power. The ability to create digital identities that grant meaningful access to opportunities is presumably comparable. But comparable how? How do we compare the power to make online identity systems to states’ authority to provide “official” documentation like passports and visas?
A better answer to this question will help us better understand how state power operates online—and how private companies compete with states over that power. It would also help us imagine (and evaluate) alternative models of identity—how they’re better and for whom.
To what extent can we define (and measure proxies of) the power that accompanies online identities & their provision? That’s an interesting question for me, one I’m likely to puzzle through in my measurement work. An answer to it would help us map who—companies, governments—wields power over what on the Internet.
Meanwhile, “decentralized” identities are emerging as a research area in computer security.These identities amount to cryptographic mechanisms that make identifiers portable, impossible for any single provider to revoke unilaterally. How do those mechanisms (re)distribute identity power—and, in practice, to whom? The answer to that question would help us assess the impact these decentralized identity systems will (and will not) have in reshifting power dynamics online.
Here are some other questions on my mind. Maybe you can answer some of them.
How do online identities relate to legal concepts of personhood—or deviate from them? What particular notions of personhood do identity systems embed—and where do those notions come from? From Western legal traditions? From somewhere else?
What are some other notions of personhood—beyond individualist, Western concepts of personhood? For example, what are some conceptions of personhood from North American or South Pacific cultures? How might those be encoded in an authentication or identity system?
Does brain-based authentication embed an alternative notion of personhood when compared to typical “login” systems? Or does it rest on the same Enlightenment notions of thinking, autonomous persons as all the others?
This post is dedicated to the memory of George Floyd.
Gray, Mary L., and Siddharth Suri. Ghost work: How to stop Silicon Valley from building a new global underclass. 2019.
Deirdre K. Mulligan and Daniel S. Griffin. Rescripting Search to Respect the Right to Truth (August 8, 2018).
Marion Fourcade and Kieran Healy. “Classification situations: Life-chances in the neoliberal era.” Accounting, Organizations and Society 38.8 (2013): 559-572.
While vastly different in the types of harms they bring, banning and “de-platforming” on social media websites touches on similar questions of access—more specifically, who gets to grant or deny access to services and opportunities. An amusing recent example: Twitter suspends Maxim journalist for allegedly impersonating ... herself?
Patrick Tucker. Refugee or Terrorist? IBM Thinks Its Software Has the Answer. Defense One. January 27, 2016.
Singh, Ranjit. “Give Me a Database and I Will Raise the Nation-State.” South Asia: Journal of South Asian Studies 42.3 (2019): 501-518.
Singh, Ranjit, and Steven J. Jackson. “From Margins to Seams: Imbrication, Inclusion, and Torque in the Aadhaar Identification Project.” CHI 2017.
Hito Steyerl. A Sea of Data: Apophenia and Pattern (Mis-)Recognition. e-flux #72 (April 2016).
My online accounts let me send a password reset to my Google-controlled email. That means my access to that account is as secure as my access to that email address.
Deepak Maram, et al. “CanDID: Can-Do Decentralized Identity with Legacy Compatibility, Sybil-Resistance, and Accountability.” Oakland 2021.
Gellner describes the “...establishment of an anonymous impersonal society, with mutually sustainable atomised individuals” as a critical facet of nations. I imagine Neuralink building mind-meshes, challenging these “atomized” models of personhood—or at least making their limitations more obvious. Perhaps this shift takes the form of a return to other traditions of personhood—indigenous traditions, perhaps?