Vélar, Ventôse, 228; Reading GANs

Oh I had lots to talk about but to be honest I'm tired an a bit sad so I'll just dump some quick thoughts on a thing and leave you to your week.

Wes sent me this to read which he's featured in of course. It's all good but I was particularly there to read How Generative Adversarial Networks (GANs) Changed the Way we Look at the World by Lenka Hámošová. I appreciate a straightforward title and it's also a useful breakdown contextualising machine learning in the story of photographic forgery and propaganda-driven editing. Interestingly, the author also pushes at the idea of over-scepticism, that as well as the possibility for social manipulation through forgery, we may become overly suspicious of images that are in fact, unaltered. In my work on this subject I've drawn many of the same parallels; Stalin, photoshop etc etc but hadn't considered the idea that publics might wholesale distrust visual media as a result of its very production. For some reason I'm minded here of the sharpie issue when Trump altered a map of Hurricane Dorian's path with a Sharpie. Despite the obvious fraud (and Trump's ongoing fractious relationship with Sharpies) there appears to be an implicit notion that a direct connection between the author (Trump) and the artefact (the adapted map) implies truth while a digitally produced image without the hand of an author visibly present could be mistaken.

To be honest, there's not much more to the text than a description of the current state of play in the technology and a call for tools to interpret images produced by GANs. It's easy to speculate on dystopian visions of total distrust and then demand better tools for verification. I'm more interested in the wider social and human effects on visual culture.