What do I look like to a machine? Computer vision algorithms “see” me as a series of boxes that can be recognized and named, a group of points whose relationship to one another predicts my emotional state, and so on.

A series of self portraits attempts to reclaim my image from this reductive algorithmic gaze, while prompting additional questions: what ways of knowing are illegible to a machine? In what ways do I want to be known, to myself and to others?

In a series of public workshops, I lead participants through my own process of analyzing the ways a computer vision algorithm “sees” my personal photo and creating an image in response. Participants discuss seeing themselves analyzed, and reflect on how different forms of marginalization impact or change the experience.

How do networked technologies create shared histories? In this installation, participants move their hands to reveal and conceal the visitors who were there before them. At the end of the session, the booth prints an oversized memento: your portrait overlaid with traces of previous users. The grainy print on fragile receipt paper provides an uncanny physical record of an ephemeral experience, at once disposable and precious.

Intimacy Chat tweaks the standard video chat interface to facilitate an awkward, playful, and tender interaction. After generating a URL from the Intimacy Chat website and sharing it with a partner, both participants’ chat windows merge in a single overlapping video feed. Together, they click through a series of intimacy and embodiment-enhancing prompts:

“Align your eyes with your partner’s eyes to establish a personal connection.”

“Exert your realistic presence to your partner.”

“Maintain a three-dimensional sense of your partner’s presence.”

The prompts’ language is adapted from a 1980s sales brochure for high end business video conferencing. I began researching the history of video chat during New York City’s Covid lockdown, when I suddenly began communicating almost exclusively over software, like Zoom, built for power-tinged business encounters. Intimacy Chat foregrounds and exaggerates the uneasy merging of closeness and control.

My first week in New York, a few months before Covid emptied the city, I took a photograph of my sister holding my niece. Later, I returned to the photograph and the memory it evoked, tracing its digital footprint through ever wider networks of data, money, and power.

In an experimental visual essay, I reflect on the relationship between a memory, its digital representation, and the far-reaching data it generates: “My memory, in the moment of its capture, was at once flattened and augmented, made both less and more.” From the color values of its pixels to its location in the world, what metadata does this intimate image carry — and what can it teach me about the drive to render our world machine readable?

With Whose Blood moves through a sparse digital space populated with stock body parts and spiky, generative text. The camera’s omniscient eye evokes a surveillance drone or control center, foregrounding and implicating 3D imaging technologies’ military origins and ongoing uses.

How much access to yourself will you allow? In this experimental narrative face filter, a disembodied voice prompts you to set up facial recognition, requesting your confirmation at each step with increasingly pushy exhortations to ‘turn your head’ or ‘smile.’ As you mime movements and emotions, your image comes more and more clearly into focus.

In March 2020, when most of the country stopped, the marketing emails kept coming. In this time capsule from the first months of the Covid-19 lockdown in the U.S., a series of AI-generated emails — trained on the corpus of increasingly frantic communications in my inbox — become a meditation on labor, corporate language, and the narrative ruptures of late stage capitalism.

For NSFW Venus’ archival interventions, I downloaded 20,000 stolen images — a sliver of the training data used by NudeNet, a nudity detection algorithm. NudeNet and its algorithmic kin use millions of scraped intimate images to train their AI models to “see” nudity, and to classify images as either pornographic or non-pornographic in response. 

From phrenology to sexology, NudeNet is only the latest imaging technology employed to sort our complex identities and behavior into rigid categories of race, gender, and obscenity. Its stolen training data constitutes a new kind of photographic archive with familiar aims — surveillance, censorship, and control. 

To NudeNet and its creators, the individuals in their datasets matter only in aggregate. NSFW Venus seeks to acknowledge and address the people in the images without reinscribing the violence of their theft. In dialogue with Stephanie Syjuco’s Shutter/Release, which uses Photoshop’s “healing brush” to remove subjects from prison mugshots in colonial archives, I digitally obscure individuals’ features, shielding their identities while foregrounding their individuality.

I trained two AI models on databases of images depicting power and privilege: the “Shirley cards” historically used to calibrate skin tones in 35mm film; and 5,000 years of carved busts from the Metropolitan Museum of Art’s collection.

The generated composite portraits move in and out of legibility, traversing the AI models’ “latent space” to reveal gaps and distortions within the data. By placing historical archives in conversation with AI training data, I explore the encoding of long-standing hierarchies into new forms of intelligence.

Sex workers were pivotal in making the early internet desirable, accessible, and profitable. But in contemporary internet histories, their criminalized and stigmatized labor is unacknowledged. Building on sex worker-led scholarship, I researched, gathered, wrote, and designed a resource for artists, designers, and technologists to engage with this history, and learn about how to support the activists leading the fight against online censorship and discrimination today.

With a launch event, I invited prominent voices in sex work activism to “talk back” to my visual research, centering their experiences and voices in a history that’s still being written.

What would a digital media project look like if sex workers were involved and compensated at every level? For Decoding Stigma’s first grant-funded project, we established a paid cohort of sex workers and accomplices to participate in a series of “communal dreaming sessions.” 

Over a series of meetings, we convened to discuss topics spanning community, safety, surveillance, and self expression, and responded to prompts with writing and visual art. An archive website collects these workshops and their outcomes, establishing a living counter-narrative of technology as collective, communal, and expressive.

In the summer of 2020, Gabriella Garcia and I started a collective.

We were both involved with the urgent, vital, sex worker-led activism and art that grew in the wake of 2018’s disastrous SESTA/FOSTA law. We were each the only advocates for sex workers rights in our respective technology-focused graduate programs, and we were tired of explaining ourselves. Inspired by Gabriella’s thesis, we dreamed of making a space where like-minded activists and researchers could meet to collaborate, organize, and simply exist, without the need to battle stigma and misconceptions.

Others felt the same way. With crucial support from Hacking//Hustling, Decoding Stigma brought together a diverse group working in and out of academia, across tech, art, design, law, public health, gender studies, anthropology, and clinical social work. What began as a series of weekly Zoom meetings solidified into an ongoing collaborative practice spanning publications, dreaming sessions, and community talks.

As Decoding Stigma’s creative director, I set out to challenge and subvert prevailing media depictions of sex work and gendered labor. In response to heavily stigmatized and stereotyped representations — and in close conversation with Gabriella’s potent theoretical writing — I plumbed dark corners of the Internet Archive to build a vocabulary of playful, defiant imagery highlighting the deep connections between erotic labor and technology.

For a residency at Headlands Center for the Arts, I engaged their quarter-century-spanning database of artists as both source material and subject. Using social mapping software and a large-scale drawing machine, I visualized a series of potential interactions between previous artists in residence.

While visitors watched, networks grew and constructed themselves, as the computer-generated visualizations were translated back into the physical space where the connections first occurred. The resulting drawings, though transcribed by machine, betray the inevitable artifacts and imperfections of their physical production.

An accompanying video essay details the methodology used to generate its weblike forms, and investigates the unexpected outcomes and ambiguities of collaborating with machines.

Commissioned for the NRW Forum AR Biennale in Düsseldorf, Germany, I’m glad you asked augments the museum’s grounds with virtually labeled park benches. Statements like “This seat is reserved for people who are missing someone” encourage spontaneous interactions between museum-goers and unaware visitors.