Thread · 6 tweets · 29 Mar 2022

"The Uncanny Gaze of the Machine," wonderful @LMSacasas as always. This triggered two related thoughts I'm still noodling over: fatalism, and the governance of inferences. theconvivialsociety.substack.com/p/the-uncanny-…

I think that the specific flavour of despair we feel at being observed by a computer is fatalism. With thin mediation (eg. binoculars) there is or will be a human to argue with. With thick mediation, the categorisation has been set in motion and there is nothing we can do.

Relatedly, many problems digital privacy are issues with the governance of inferences. In the specific (fascinating) research covered here, people carefully curated their online presence, are comfortable with employers checking it, but not using computerised means.

When we curate how we present, we know that others can draw further inferences. But there are rules to that: it's limited by human cognition, you can often appeal a mischaracterisation by talking to the person, you know how to course-correct... That's informal governance.

A computer inference however may draw upon massive data and complex correlations, you can't argue with the machine, and you can't see how to change. You have no say and governance is replaced with management. It's like data informed nudges vs manual ones.

There's a lot more to say on all this, but thanks for the inspiration @LMSacasas!