In Foam Magazine #66: MISSING MIRROR,
me and others look at the growing overlaps between art, technology and society,
exploring how the recent advancements in AI impact our relationship with the image,
ourselves and our perception of reality.

How do we form a truthful image of the world when credibility is questioned?
And vice versa, how do we recognise ourselves in the images around us?⁠

I contributed with a portfolio text on Leda Sadotti’s project ‘Beautified-AI’

Beautified-AI

The world is rapidly advancing in digitalisation and technological growth. This digital transformation presents a paradox: it offers global connectivity while also fostering deep divisions and misunderstandings. In navigating this terrain, artists emerge as crucial pioneers and investigators. They have the means to merge imagination and technology, creating spaces for experimentation that pave the way for novel future possibilities. Art, in that regard, has the power to inspect, communicate, and disrupt on profound levels. In this context, let’s take a closer look at Leda Sadotti’s project Beautified-AI, which debates the issues of synthetic artificial intelligence (AI) imaging and uses augmented reality (AR) technology to politicise fabricated profiles.

Beautified-AI began as a study into catfishing online profiles and debating the rights of AI-generated forms, stemming from conversations Sadotti had with a GDPR (general data protection regulation) and wider data protection lawyer, through which she came to know more about the impact of image fabrication on digital autonomy and its threat, especially the broader implications of our online footprints. This sparked further research into practices of synthetic imaging technologies, where the method unfolds in distinct stages and is intentionally vague and complex.

Having previously worked with 3D modelling and creative programming, Sadotti had not considered the capabilities of AR technology in forming communal spaces and digital archiving prior to this project. When asked by OpenSpace (an AR initiative comprising artist commissions and a mentoring programme for young artists) to consider the limitations of how extended reality (XR) is showcased within creative spaces, Sadotti began to consider the ability of AR to construct a diverse community. Whereas she first felt constrained by the technology — since she saw it as an observation tool — this feeling shifted when realising the technology could facilitate encrypting a ‘natural face’.

By collecting all types of fake profiles, from categories labelled as ‘beautiful’ and ‘natural’, Sadotti made her own machine learning model. Letting the model run, it created distortions that the artist had no control over, and this was exactly what she wanted to explore — seeing how she could play around with a fixed dataset and encrypt it even further, because, none of the faces were actually real in the first place and the distortion created another layer of ‘fakeness’.

The next step in the process was figuring out how those ‘people’ could be politicised, by being shared and then letting the faces themselves become something. Through machine analysis, Sadotti explored the livelihood of these synthetic beings within a series of speculative processes. She slightly changed the original model so it would be able to generate text, and trained it on vector graphs formed by using web pages in response to the project and to poetry Sadotti wrote off the initial model. She purposely used words like ‘beautified’, which are common on platforms enabling companies to buy fabricated headshots — these words aim to humanise synthetic faces and were used to produce the natural language processor colours for Sadotti’s profiles.

After all the models ran their course, an AI-generated community of profiles was born. Instead of having used technologies as commodities, Sadotti recognised their profound significance beyond mere (commercial) utility. By having formed complex, digitally woven ‘live beings’ from AI-generated portraits with AR distortion, she speculates on the impact of fabricated internet user profiling and examines its development, which, as it quite unexpectedly turns out, presents a bias towards ‘attractive’ features. With Beautified-AI, Sadotti creates a discourse around this discrepancy by categorising, labelling, and ‘enhancing’ images that are not from actual human beings but are treated as such nonetheless.

When art questions, it often doesn’t come up with answers. Rather, it poses more questions through the extensive analysis of concepts that are far from easily defined and encapsulated. The time we live in is exciting, though confusing. Technology is impacting our society in an accelerating way while our way of thinking is not suited to keep up with it. This leaves us in a void, so we look to artists for guidance. Leda Sadotti delves into the void, investigating the societal implications of emerging technologies and how advancements in artificial intelligence alter our interaction with image prompts, a deeper reflection on self-perception, and the nature of (un)reality. What reflection meets your eyes in the mirror? Beauty, authenticity, or perhaps a stranger shaped by the world’s expectations?