Class of 2022: Connie Stewart – Machine learning

Using vintage photographs and embroidery, Gray’s School of Art student Connie Stewart explores the pervasive nature of big data and AI. By JAMIE LIMOND

A group portrait. Junior sports team, rackets and blazers. Sepia-toned. We sense the camera freaking out as it tries to apply its face recognition software to the 60 or so kids in front of the clubhouse, their faces partially obscured by overlapping rectangles. But we shouldn’t automatically trust this image. It obviously pre-dates this kind of technology. The frames aren’t part of the image but are physically embroidered onto it. The intelligence behind them is human, rather than machine.

The portrait is just one of over a hundred similar photographs assembled by Gray’s School of Art student Connie Stewart. For the past couple of years, she’s been researching contemporary surveillance practices and the evolving field of face recognition. “We’re aware that we’re being observed, but I’m interested in how these observations are implemented in the world, particularly in social networks,” she tells me over Zoom. “Ideas of Big Brother, CCTV – we’re long passed all that. I’m more interested in looking into present concerns over big data and the use of AI and machine learning.”

We’re pretty used to seeing face recognition boxes pop up on our camera phones, but the physicality of Stewart’s embroidery makes them impossible to gloss over. In some images the structure of the face is broken up into segments and planes, almost like Renaissance instructional etchings. “For me it’s been a bit like life drawing,” Stewart says, “going back to learning the anatomy of the face.”

Connie Stewart, Recognition Portrait No.24, 2022, embroidered photograph

The images she works with are mostly formally posed, anonymous. So anonymous that you feel like you’ve seen them before. They play games with our own very human recognition software, sitting somewhere between the familiar and unfamiliar. “They used to tell people not to smile in photos because it made them look silly, so they look nothing like the photos we take today. They feel almost otherworldly. But the more you look at them the more you remember these people had lives. And you wonder, ‘What would they think of me doing this?’”

For some viewers the act of embroidering feels like a kind of violation. “They see it as form of vandalism, as a violent act. And surveillance practices, especially when they’re non-consensual, are in a sense quite violent. The photos are available online because no one wants them anymore, but I always ask the person giving them away for their consent.”

The non-consensual use of likenesses is also a major concern with the deepfake, a technology which for Stewart seems to have outpaced the ethical conversation. “These processes develop faster than we can keep up with. There’s so much potential for their creative use, but of course they’re put towards unethical ends. It’s important to remember it’s not the fault of the AI – it’s the fault of the people who programme it. The machine is really just the thing between the humans on either side.”

Connie Stewart Motion Detection No.1, 2021, embroidery on vintage photograph

In that sense, the technology is a mediator, like any work of art. And AI wouldn’t be the first new medium accused of being coldly analytic, inhuman. Artistic practice also relies on a form of pattern recognition, the ability to put a frame around something, focus in, abstract. I wonder if the process has brought Stewart and machine to some closer level of understanding.

“There’s an element of understanding,” she considers. “At first there’s a connection with the photographs, there’s a sense of, ‘these are real people, maybe I need to be careful’. But it’s almost become a machine process. A lot of love and care still goes in, but even the way I embroider has changed over time. Initially I handled them very delicately, trying not to tear the edges, it was an intimate and caring process. But as I made more of them it became more fast-paced, and the connections became less prominent.”

The embroideries will be a major part of the degree show, but Stewart also plans to incorporate an interactive digital work. “I want something the audience can interact with directly. I don’t know if I want to say too much about it at the moment because I like an element of surprise in my work. But I do want to push the potential understanding between humans and machines, which I’m really excited to look into in the future.”

Degree show: 11-18 June 2022, Gray’s School of Art, Garthdee Road, Aberdeen. conniestewart.net

Main image: Connie Stewart, Sensitive Data (detail), 2021, embroidery on vintage photograph

For more Class of 2022 features, Q&As and profiles, read the a-n Degree Shows Guide 2022:

https://issuu.com/anartistsinfo/docs/a-n_degree_shows_guide_2022