Perverse affordances hails from 2018, which from today's era feels like machine learning deep history.

Starting with a bot that crawled large global social media platforms and amassed a custom dataset of over 10000 screenshots, Friend then used them to train a custom machine learning model. As a final step, Perverse Affordances used the new “social media screenshots” that are generated like wireframes, to imagine new online interactions or possible web layouts.

Taking social media as a location - a place where machine learning algorithms are "watching" our online selves all the time, Perverse Affordances turns the act of "watching" around on itself. How does an algorithm sees an interface? Can we make the platform visible by looking at it with a new (machine) eye?

Sarah Friend is an artist and software developer from Canada and currently based in Berlin, Germany. In 2023, she was a research fellow at Summer of Protocols, led by Venkatesh Rao and the Ethereum Foundation, and in 2022, she was a professor of blockchain art at the Cooper Union. She has exhibited at and worked with MoMA (NYC), Centre Pompidou (Metz), Kunsthaus Zürich, HEK (Basel), Haus der Kunst (Munich), ArtScience Museum (Singapore), bitforms (NYC), Albright Knox Museum (Buffalo), Rhizome (NYC) and KW Institute for Contemporary Art (Berlin) among others.