Te-Mao Lee’s In-Between beings is currently on view at IA&A at Hillyer, presented with support from the Ministry of Culture (MOC) of Taiwan and the Taipei Economic and Cultural Representative Office in the United States (TECRO). The exhibition consists of four works, each centered on underwater environments. Lee’s turn to the sea reflects a conceptual interest in a vertical sense of time, spurred by recognition of the way that tectonic forces along Taiwan’s east coast have gradually raised remnants of prehistoric coastal sites from the ocean’s floor. The artist’s work focuses on such sea beds in an attempt to, in his words, “connect the present with a future two thousand years ahead”—one in which the now-submerged grounds have themselves risen to the surface.
For Lee, the nature of that distant future seems to hinge, at least in part, on the evolving relationship between humans and AI. Not only do individual works explore the relative strengths and vulnerabilities of each, but so, too, does the show’s organization as a whole. Two video works are built around the deployment of AI, while the other two foreground the experiences of human scuba divers (including the artist). That the conditions of possibility for the exhibition itself turn out to be both AI and human action poignantly signals our current state of “in-between,” cued from the outset by the exhibition’s title.
Te-Mao Lee, The Misfit Names, 2025, installation view. Two-channel video, real-time object detection system, large language model. Photograph by Katherine Markoski for East City Art.
The Misfit Names (2025) and Where Thought Never Lands (2025) are the two works that engage directly with AI. In the former, a large projection displays complex aquatic topography with marine life drifting in and out of view. A smaller screen on the adjacent wall then shows the same visual field being analyzed by an AI that has never encountered underwater worlds. Rectangles surround items designated for identification; above them appear the object detection system’s deductions, while descriptive sentences produced by a large language model run across the screen’s bottom edge.
Te-Mao Lee, The Misfit Names, 2025. Two-channel video, real-time object detection system, large language model. Courtesy of the artist.
If these superimposed geometries visualize technological aspirations to rationalize the scene, then the often humorously off-base identifications—a fish is a bird—register the limits of that ambition. But, as the juxtaposition of the video streams suggests, Lee’s message is more nuanced than this simple critique. Seen side by side, the moving images invite reflection on what is lost when we attempt to box in the world. It hardly seems incidental that one video is projected while the other is contained by a monitor. The large-scale video, with its undisturbed visual fluidity—fluctuations in the water, dynamic aquatic life—implies that understanding, and even beauty, resides in a relational openness, a fluid whole.
Te-Mao Lee, Where Thought Never Lands, 2025, installation view. Single-channel video, large language model, DoS audio data transmission. Photograph by Katherine Markoski for East City Art.
At the same time, it is hard not to see the work’s “misfit” language as taking on a certain poetic cast, and this quasi-surrealist sensibility also pervades Where Thought Never Lands. The question of what lies beyond consciousness is crucially at stake for Lee, much as it was for the historical surrealists. Today, one answer increasingly points to AI. In this work, the camera slowly moves across an underwater landscape. The large language model Ollama then watches this footage and, in response, generates a series of statements that are presented as subtitles. Because the language model operates in a closed system, its questions are repeated and refined as the video loops. Swings from formulations like “Is memory simply rearranged data now?” to “The blue consumes, a comforting decree” are echoed by the camera’s uneven traversal of the seabed’s ridges and by the dissonant audio of the sound frequencies used to transmit Ollama’s language. According to the label, Lee is interested in whether the digital thought on view could evolve into a higher form of consciousness if given enough time and space. While the uncertainty surrounding this query—which might be to say, surrounding the future of AI—is borne out by the haziness of the underwater footage, for this viewer, the enduring poetry of the ocean floor—its undulating folds like so many stanzas—came to feel certain.
Te-Mao Lee, The Unreached Surface, 2025. Single-channel video, glass, ecological bottle. Courtesy of the artist.
The watery depths are central again in The Unreached Surface (2025), which consists of a large-scale projection, glass, and what the label describes as an ecological bottle. In the single-channel video, Lee and two other divers descend twenty meters below the surface to collect non-cohesive sand before rising again. Off-screen, this sand is melted at high temperatures to create glass, which is then blown into sealed bottles, each holding a just visible ecosystem of plants, water, and air. One of these bottles and pieces of glass are placed on pedestals in the gallery. Lee’s statement makes his intent clear:
The work constructs a space of viewing, approaching, and loss–echoing
Deleuze’s idea of the fold-crystal: a progression of time, crystallized into an
experience that is perceivable yet impenetrable.
Overall, Lee’s work is about transformation, surely, but alchemy feels the better term. Humans are beings of the sea, able to penetrate its depths and emerge again. Sand is not just turned into glass, it comes to foster a self-sustaining world. But, importantly, the limits of human magic are also on view. That glassed-in world is visible but, unless broken, inaccessible. The reverent steadiness of the divers’ journeys downwards, largely tracked through their bubbles, is disrupted when they appear near the marine floor, wielding flashlights and pickaxes, jarring presences in this unfamiliar terrain. As they rise again with their bag of sand, the reddish buoy they hold may read both as a marker of their ascent and as a warning about the future in which the whole ocean’s ground has surfaced. Whether the humans are the ones sounding the alarm or the cause of the danger is a question Lee seems to deliberately leave unresolved.
Te-Mao Lee, 45 Minutes, 2025. Single-channel video. Courtesy of the artist.
The strongest and most captivating work in the exhibition is 45 Minutes, also the purest manifestation of Lee’s interest in time. This single-channel video was recorded as Lee and other divers hovered twenty meters below sea level for forty-five minutes, the longest time a scuba diver can safely remain at that depth. The result is a moving portrait of water. At times, the ocean’s colors form painterly swaths across the screen. At others, the water’s visual rhythms are punctuated by arrays of exhaled bubbles and their sounds, which serve as the work’s audio component. These bubbles, in both audio and visual form, register the divers’ presence and thereby subtly interweave the beauty of the work with that of human life. Similarly, as the bubbles animate the water, they remind us that the ocean itself is a living being, replete with its own wonders and complexities. Ultimately, it is tempting to draw a connection between the divers pushed to their limits and an underwater ecosystem equally under threat. (A line from the adjacent Where Thought Never Lands kept coming to mind: “The algorithms hum, predicting my decay.”) As Lee suggests throughout this exhibition, we are living in a state of in-between: will we rise up—or run out of air?
Te-Mao Lee, In-Between beings at IA&A at Hillyer, 9 Hillyer Court NW, Washington, DC, 20008, September 6-28, 2025. Hours: Tuesday-Friday, 12 to 6pm; Saturday-Sunday, 12 to 5pm. athillyer.org.
