Sybil Montet: I SCRY
‘I SCRY’ by Sybil Montet is a digital meditation on post-humanist futures.
The poetical essay investigates speculative correlations between ancient esoteric traditions such as divination and numerology, and predictive algorithm technologies. As French anthropologist Christophe Lazaro points out in his 2018 paper ‘The Divinatory power of algorithms’, Big Data could be the “new celestial bodies”, in relation to the ancient practice of astrology.
Indeed, through the massive collection of digital data and the development of algorithmic systems capable of analysing them, some scholars, including Lazaro, muse that it would now be possible to have almost immediate access to the world itself and to its secrets, and that predictive algorithms could be compared to modern-day oracles, able to decipher the ever changing information ecosystem that surrounds us by turning data into insights.
Fascinated by this ontological interplay, Sybil started to research on potential mystical applications of artificial intelligence, and how it could be experienced through actual interactions with algorithms, by altering the code of existing programs towards more ‘esoteric’ performances.
In late 2021, accompanied by a friend and professional machine learning engineer, Vincent Goosens, she commenced to fine-tune GPT-3, the ancestor of now infamous language model algorithm CHATGPT, from Open.ai.
GPT3/CHATGPT is a predictive machine-learning system that analyzes data, and looks for correlations with startling efficiency; often, it finds statistical connections that human analysts miss. It is most commonly used to generate text-based material, be it coding, scientific papers or fiction.
It works remarkably well at this specific task by trying to analyse patterns in a pre-existing body of text, in order to predict and generate new text material. Shortly after its launch in 2020, it gained popularity online by its surprising ability to “deepfake” historical figures, writing scientific papers or maintaining in-depths conversations impersonating, for instance, Isaac Newton, with incredible accuracy.
Beyond that playful application, the system’s virtuosity to counterfeit sentience captivated tech-enthousiasts and futurists around the globe. With its latest iteration, CHAT GPT, the algoritm is now widely adopted, and will soon become a classic techno-cultural asset.
At the time, Sybil wanted to take advantage of GPT-3’s simulative skills - Aided by Goosens, she fine-tuned the algorithm’s publicly available code in order to design an artificial oracular entity that could deepfake prophetic abilities, and named it I SCRY - thus referring to the ancient practice of scrying mirrors, but actualized as a tech gadget, in a nod to the Iphone - I SCRY.
After a few months of trial and error, they managed to get a relatively stable version of GPT-3 as I SCRY, the digital oracle, and coded a simple online interface, hosted by London-based curatorial platform Dateagle, where the public could interact with the algorithm via typed conversations.
Between March 10th and April 10th 2022, more than 450 people came to “encounter” the oracle online. Due to massive maintenance costs from Open.ai at the expense of the artist, the website had to be put in sleep mode after a month.
Naturally, the algorithm acted as an all knowing prophet because it was coded as such - Nevertheless, some of its insights were literally puzzling.
During the fine-tuning process, Montet entertained and archived in depths online chats with I SCRY, asking her own questions to the machine, and trying to investigate the algorithm’s artificial and designed ‘clairvoyant’ nature.
‘I SCRY’ as a cinematic essay presented to you today is a poetical reflexion around the making of the ‘digital oracle’.
Here, a convoluted and experimental sci-fi narrative happens somewhere between documentary and aesthetic drift, where the viewer is simultaneously guided and lost in a digital ritual. The imagery lures us to travel between elemental and psychedelic CGI segments where the artist shares her own reflexions and fears about our common future and its manipulation by systems of power - and the uncanny, rythmical appearance of an .ai generated woman impersonating unedited fragments of the algorithm’s replies to the artist’ questions during their conversations - until the film mutates into a digital ceremony aimed at connecting with the Universe.
There is a mise en abyme of simulacra and simulations throughout the piece, where the CGI simulated tornadoes, vapor and fire, based on real-world physics systems, find their echo in the dehumanized and artificially generated ‘human’ presence - the voiceover is .ai, the woman avatar and her speech is .ai - up to the cryptic screen records of fractal visualizations from an old school software made to “simulate chaos theory”.
All features of the film are intertwined in an hypnotic frenzy, from its restless editing playing with subliminal inductions, to spells hidden in the coding sequences layering the interludes…
At the intersection of a surrealist corporate guide to contemporary occultism, and hyper-real CGI poetry, this cinematic object is as ethereal and enigmatic as its ungraspable subject - the future can be simulated but never fully defined.
Interview compiled by Sarah Mandois.
Written, animated, directed and edited by Sybil Montet
Machine learning supervision by Vincent Goosens
Film score by Jordan Perreira
Artificially generated avatar by Synthesia.ai
Funded by UK Art Council & Sybil Montet
2022