Is it true ? The Post-truth Archive Factory
Exhibition at The Briggait in Glasgow - 7th December 2021 / 6th January 2022
‘Is it true? The Post-truth Archive Factory’ is the result of an entanglement between archives and artificial intelligence exploring our relationship to the truth and its counterfeits? This project presents a triptych of artworks which deliberately blur the boundaries between true and false.
By using archives considered to be the ancestors of the tabloids – which have therefore created a complex relationship with reality – Marion Carré generated false textual archives using artificial intelligence. In producing a form of 'mise en abyme' of the fictitious, she created artworks using authentic and generated archives to question the conditions of their production, perception and circulation.
Perceived as absolute and unchanging, together the archive and the truth make us forget their complexity and the constructions from which they originate. In the context of liminality between what is true and what is false and the continuous flood of information, we sanctify the truth while participating in the spread of false information in its quest. These forged truths become proof to build our own, freed from reality. If our individual memories are corruptible, how immune is our collective memory to false information?
This exhibition is a triptych of 3 artworks exploring the subject : Living organism, Selective memories and The Post-truth Archive.
Living organism
Installation : Marion Carré
Video : Oscar Boissiere
Music : Paul Schmite
Hosting place : Library of the École Nationale des Chartes (Paris, France)
During the exhibition, the originals paper rolls from the installation and the video of the installation were exhibited.
Installed on the high bookshelves of the École Nationale des Chartes, surrounded by old books and archives, printers copy on long paper rolls the texts of original Broadsides and AI generated ones.
The mother archives are texts from the Broadsides archive of the National Library of Scotland. Known as the 'ancestors' of the Tabloids, these archives have a complex relationship with the truth. The selected texts are either text recounting trials, which present here a form of ‘mise en abyme’ of the relationship to reality, or texts that relate very surprising facts to the point of doubting that they really took place. These texts rarely leave us unmoved because they carry an emotional charge that further disturbs our appreciation of past events by establishing a particular connection with them. They were used to train the artificial intelligence algorithm that generated the fictitious texts of the daughter archives.
Here, the artificial intelligence endows the archive with a reproductive function. Through the process of generation, the mother (original) archive creates daughter (generated) fictitious archives. This process underlines the living character of the archive as well as of the truth that we perceive as fixed whereas they can evolve with time but also the interpretations that we make of them.
Through generation, artificial intelligence defies the finitude of the archive. Nourished by original Broadsides archives, it randomly composes new texts that are added to the old ones, thus leading to an exponential and limitless accumulation process. Does the current rush of traces and the exponential growth of archives have anything to do with the presentiment of our own imminent disappearance? At a time when we delegate more and more tasks to machines to the point of attributing human characteristics such as intelligence, could we ask them to continue writing History without us? How far can we delegate to machines?
Selective memories
Selective Memories is an online participative artwork that proposes to its participants to choose, for each text which is presented to them, if it deserves to be archived or destroyed according to them.
All the texts are displayed randomly and presented on the same plane, in an identical format, thus reproducing the phenomenon of loss of reference points that social networks create with regard to the information they relay. In order to respond to the injunction given to them, users must still find indicators to decide which texts to keep or destroy. Will verisimilitude be the only criteria? Will the emotion that certain texts can provoke get in the way? Can 'fake' texts manage to produce such emotions in their readers? What place does emotion have in our apprehension of the true and the false?
Selective Memories calls upon a reflex that we use every day in our use of social networks: to separate the true from the false in the information that is presented. However this work, often unconscious, is not without consequences for our memory in which can slip false information to the point of developing false memories. In this case, the mechanism is reversed: the participants think they are consulting archival texts whose format does not allow them to identify at first glance which ones are authentic or not. It is only after having indicated their choice to archive or destroy a text that they discover the existence of this parameter which can potentially call into question its validity.
To participate : selectivememories.org
The Post-truth Archive
The Post-truth Archive is the output of collaborative work made by the participants of Selective memories online platform. The red register gathers all the texts that participants decided to archive on the platform. It is surrounded - crumpled and scattered on the ground - by texts people decided to erase. It blurs the line of truth and fake because original and generated broadsides can be found on the register and laying on the ground.
It explores the perception mechanism that allows us to give credit to fake and propagate it. Is crowdsourcing reliable in this situation ? The approval of these texts by other peers can potentially introduce a confirmation bias in the viewers. The way things are presented can also have an impact and be a source of bias, and this is why the register is using the codes of how official archives documents can be presented.
Furthermore, the very status to be given to this body of text is questionable. It could be considered as corrupted as it would mix truth and falsehood. Yet the treatment of these texts may have even more to teach us. Is this corpus only a reflection of the true/false cleavage? Does the fact that some false archives were retained by certain participants give them a special status? Can the value given to this corpus go beyond the question of its veracity?