Aral Sea Ecocide
The ‘Aral Sea Ecocide’ presents a series of AI guided images and videos depicting landscape mutations of the Aral Sea crisis. Straddling the border between southern Kazakhstan and northern Uzbekistan, the Aral Sea was once deemed the fourth largest lake in the world. However, since 1960s, the Soviet irrigation project and its classified biological weapons facility in Vozrozhdeniya Island, led to one of the most notorious ecological catastrophes.
A curated selection of 300 satellite datasets spanning from 1964 to 2023, were manually collected from the NASA's Earth Observing System Data and Information System (EOSDIS) Worldview. The obtained datasets illustrate the following impacts: land surface temperature, aerosol index and drought hazard frequency and distribution through their varying colour tones. At the same time, there is an incomplete picture of the effects of bioweapons on ecosystems and public health. The design process harnesses the blindspots and discrepancies in the datasets as a catalyst for training machine learning models. Emerging from this is a complex feedback loop between the designer and the machine, which seeks to disrupt the precision-driven and fine-tuning purposes of machine learning. A fictional narrative provides a set of parameters guiding the algorithm, where each generated output becomes the new input data for subsequent training. In this iterative approach, shapes and patterns start to deviate from its original reference, amalgamating into a slow violence. The surface of the Aral Sea is like a palimpsest, a document undergoing constant erasure, yet whose traces linger across spatio-temporal scales.
Generative AI (LLMs, Stable Diffusion)
Data-driven visualisations, animation