Logo

Riksdagen
The Parliament

An art project by Andrea Nordwall

View project on GitHub machinesarenotpeople/riksdagen

The project is maintained by machinesarenotpeople

Riksdagen (The Parliament) is an custom AI that generates photorealistic portraits of Swedish members of parliament. It is a custom machine learning model trained on about 1400 press photos1 of all Swedish members of parliament from 1992-2022.

The process has imprinted the model with detailed visual information on the faces of power in Sweden during this period. As the generated images of the machine are based on properties it has learnt from the material, the output becomes a statistically accurate representation, prediction and reflection on who we choose to govern us, as well as a comment on algorithmic self-reproduction of power in politics.

Physical Outcomes

The work is intended to be shown as a series of 349 framed portraits on a wall, placed in the same formation as the seating in the Swedish parliament chamber plenisalen. In front of the wall, a screen showing a latent walk video of generated parliament members turning into new parliament members. The 349 generated members of parliament were weighted on gender and age statistics of the Swedish parliament as of August 20222.

Output from Riksdagen.

Riksdagen (the Parliament) was shown at the AI art exhibition Sentient Machines? in August 2022 at Färgfabriken, which was part of the Futureless Festival for Tachnology and the Arts with support från Kulturrådet (Swedish Arts Council) and Nordisk Kulturfond (Nordic Culture Fund), curated by Paola Torres Núñez del Prado.

Art Book

The work has been published as a book with a print run of 100 copies, documenting the 349 generated members of parliament. Contact andrea dot nordwall at gmail dot com if you are interested in acquiring a copy of the book.

Software Architecture and Model Release

The software architecture of this work (StyleGAN23) is based on research by Karras, T. et.al. (2019)4.

The AI model Riksdagen is made available (here) as a digital art work for personal use.

All portraits on this page were generated by Riksdagen.

  1. Foto: Sveriges riksdag 

  2. https://www.riksdagen.se/sv/dokument-och-lagar/riksdagens-oppna-data/ 

  3. https://github.com/NVlabs/stylegan2 

  4. Karras, T., Laine, S., Aittala, M., Hellsten, J., Lehtinen, J., & Aila, T. (2019). Analyzing and improving the image quality of StyleGAN.