A Software Engineer's response to Elite Capture and Epistemic Deference
It always a great feeling when you read something so good it inspires you to do some writing of your own. I recently read Being-in-the-Room Privilege: Elite Capture and Epistemic Deference, by Olufemi Taiwo. If you haven't read it (which I highly recommend doing), I will try to briefly condense some of its key points and then elaborate on my thoughts and where to take things from there.
If you work in tech or any field with a social justice bend, you will have likely heard from well-meaning colleagues, "Well, as a white man, I have no idea what your experience as a Black/Asian/Latino/Indigenous man/woman/nonbinary person, so I defer to you in this area." A primary concern the author raises with this norm, described as deferential standpoint epistemology, is that it is highly susceptible to "elite capture". Elite capture is the process in which the most elite members of a marginalized group commandeer the policy agenda and resources of the broader minority community in service of the elite agenda. The interlocking system of feedback loops and filtering that place us where we fall in society, and which underprivileged individuals must contort through, guarantees that when a minority is in a room of power, they're much more privileged than most in their identity-group. For a real world example of this, see 50 Cent and Lil Wayne's recent political endorsements.
Taiwo then suggests an alternative, called constructive standpoint epistemology. The key differentiator is that rather than deferring to "elite tokens", can we simply take account of what our specific goals and objectives are? Can we build institutions around these goals, with accountability and transparency baked in?
As a practitioner within the "AI Infrastructure" space at your local Big Tech Co, this is something I spend a great deal of time thinking about, albeit from an engineering perspective. I'll elaborate on three main concerns that I feel must be addressed in any constructive standpoint epistemological approach to organizational governance:
What are the right goals
The task of selecting the specific goals and end results is the most obvious challenge in for constructive standpoint epistemology, but also potentially the hardest to deal with. This is where the deferential approach is most tempting, as you can simply rely on your tokens to decide on the goals for you. Great care will have to be taken to establish a process in which specific needs and priorities can be solicited from the marginalized groups themselves. Significant amounts of work will have to go towards trust building and coordinating amongst underrepresented people. We don't have to reinvent the wheel here, however, and can look to existing work within the deliberative democracy space as a starting point.
How do you measure the right goals?
Once you've decided on that your key objectives are, the next question is, how do you measure whether you are making progress? What are the metrics through which the institution is held accountable and how are these metrics calculated? In the industry, we refer to this as measuring success in the right currency. What is the right currency? The one that most impacts the client. If our "client" is marginalized people, the right currency is directly tied to their lived experience and quality of life. This sounds like a basic aspect of public policy, but is often the least discussed in popular media.
For something like addressing Flint's water crisis, this question of "right currency" has a simple answer: the amount of lead in the drinking water. For other policy initiatives, the answer is less clear. For example, if we wish to implement universal health care, what are our metrics for determining whether people's live are materially improving under such a system? It's not hard to imagine a decrepit, Comcast-style health care monopoly that provides universally terrible patient service. Likewise, if we wish to abolish the police, what concrete metrics do we hope to move by doing so and how are these metrics tied to the lived experience of those suffering under the carceral state?
What is the process for iterating on those goals?
It's unlikely that the first pass at any initiative concerned with social justice will be a complete slam dunk. It's likely that one will have to go back to the drawing board, reconfiguring the key goals and the approach to measuring them. However, it is essential that there is a process in place for doing this reflection and iteration. The United States has a fraught history with trying bold social welfare programs, only to divest or outright cripple them when initial success isn't all but certain. Thus, defining a process for iteration can't be an afterthought, it must be baked into the governance strategy itself.
The good news is, all these problems should be incredibly familiar to us folks in tech land. Anyone that’s done high level work on a technical project knows the process of determining, measuring, and iterating on your objectives and key results. The details of this process almost certainly won't map 1:1 with addressing social problems, but I believe much of the spirit of the endeavor will.
In essence, the tools are out there, we just gotta use them.