Exhibitions
TIFFANY CALVERT: SIGNAL TO NOISE
Tinney Contemporary / 237 Rep. John Lewis Way N. November 7 - December 23rd
Tiffany Calvert
TINNEY is proud to present Signal to Noise, a solo exhibition featuring works by Tiffany Calvert. The exhibition will be on display from November 7 through December 23, 2026. The opening reception will be Saturday, November 14, from 6 to 9 PM, in conjunction with the second Saturday Art Crawl.
Calvert’s prescient studio practice has been preoccupied with machine learning and image generation for the better part of a decade; her layered and glitched floral still-life paintings are only gaining more relevance as AI encroaches on every facet of modern life. Calvert’s process begins with a dataset of 650 Dutch Golden Age still life paintings. These meticulously detailed floral works gained prominence during the Tulip mania of the 1600s, now regarded as an early financial bubble.
These images are used to train a diffusion-model AI developed in collaboration with computer science researchers Ulubek Kamilov (UW Madison), Shirin Shoushtari (Computational Imaging Group, McKelvey School, WashU) and Chad Henry (Director of Research and Innovation, Sam Fox School at WashU). This software is the subject of concurrent research for its usefulness in preventative medicine, medical imaging, and related fields.
Calvert prompts the software to generate a new image based on her custom dataset. Diffusion models interpolate noise into training data until static overtakes the sample, then reverse this process to generate a new image. Whereas the now widely available text-to-image AI platforms use datasets of billions of images (scraped from the internet without concern for the inherent biases ingested through this data, which is then kept secret under the auspices of trade secrets) Calvert’s model uses a comparatively minuscule, ethical dataset of images taken from the public domain. The generated images mimic the palette and general composition of the sample images, but uncanny deformities and glitches bely their origins.
Calvert chooses from these novel images and prints them to canvas. She then applies cut-vinyl masks, which are patterned after Cubist paintings, to the surface of the printed image. Calvert then applies oil paint over the unmasked areas in thick, gestural strokes. These painterly interventions in the ai image are a methodology for investigating the limitations of machine learning. Still-life paintings have particular resonance in Calvert’s inquiry for their art historical usage in responding to technological innovations—from the memento mori of golden age florals, to modernist fracturing of space in response to industrial alienation, to dada’s mutated figurative corollary to the ravages of WWI. In line with this tradition, Calvert’s abstract interruptions can be read as an identification with noise as evasive manouver—in an era where surveillance data is collated and weaponized against a human population by AI software, such a tactic becomes necessary. At the same time, abstraction serves as a metaphor for the reductive language of AI image classification, which reproduces the programmers' biases without fail.
