16 October 2020

Seminar Q&A

I gave a short talk recently about some of the stuff I’ve been working on (I’ll share the PDF/video sometime soon). After the talk, someone asked a question that, since I didn’t copy the question down to preserve their anonymity, I’ll summarize/restate a bit: Given the issues with metrics describing our lives, should we deal with the inherent shortcomings of algorithmic systems, or should we work on developing more metrics and measures? or a bit of both?

This is a great question, and it makes me think of a point David Graeber made in The Utopia of Rules, where he talked about Jo Freeman’s essay “The Tyranny of Structurelessness”, and how cliques emerge, even in egalitarian, feminist, anti-authoritarian groups. the idea is that as groups grow, these cliques become more influential and powerful, and that’s obviously not in the spirit of egalitarianism.

Graeber points out that there are two ways of addressing that: one is to make the cliques into formal committees so that their unequal influence at least comes under rules that foster egalitarianism.

But he points out that this essentially legitimizes that power imbalance. He asks why we don’t just dismantle the power structure itself.

That’s all a bit of a story to get to say that I think of the development of metrics and measures as a way of formalizing the inadequacy of metrics and measures to describe our lives. There are histories and traditions that shape our lives in all sorts of ways that are hard to talk about, let alone measure. And I’d argue that - no matter how many dimensions we describe - there’s virtually a limitlessly more complex real world not being described by those metrics.

I think, like Graeber, that we should tackle injustice directly, rather than formalizing a set of rules to constrain the injustice. Often that means looking a whole domain of problems, or at the people being dealt with by AIs, and deciding that AI just isn’t up to the task because they just can’t “think” deeply about our lives the way we can (emphasis on “can/can’t”; Graeber also writes about how and why bureaucracies do such stupid, absurd, harmful things to people, so “put people in the mix” is hardly a fix for everything).

I think this is a reasonable tension to experience, and while I obviously have a position on it, I recognize that someone with the other answer isn’t necessarily acting in bad-faith (although, like, I still think they’re wrong lol).

If you have something to say about this post, email me or tweet at me.