Title：The Algorithmic Gaze: Understanding AI Through Art
Date：2025/6/6 13:40-14:50
Location：R101, CSIE
Speakers：Tom White, Victoria University of Wellington
Host：洪一平教授
Abstract:
Artist-researcher Tom White explores how art can reveal the inner workings of artificial intelligence, drawing from his work at the intersection of visual art and machine learning over the past eight years. White will present the evolution of his artistic practice through three phases, each illuminating different dimensions of AI perception.
The first phase examines the fundamental differences between human and machine perception through computer vision systems. White creates abstract visual compositions that appear innocuous to human viewers yet are consistently flagged as inappropriate by AI content filters—revealing surprising misalignments between how we see and how machines interpret visual information.
In his second phase, White investigates models that integrate language understanding with visual processing. This work explores how AI systems develop knowledge of the world through multimodal learning, transforming their "Algorithmic Gaze" from rigid classification systems to more fluid interpreters capable of leveraging the full power of language to constrain their operations.
White's most recent work employs mechanistic interpretability techniques to explore neural networks' abstract latent spaces on their own terms. Rather than imposing human interpretations, his approach asks: "What do these systems have to teach us?" By visualizing neural network features directly, White creates art that functions as a translation between machine perception and human understanding, offering new perspectives on both.
Through these artistic investigations, White invites us to reconsider our relationship with AI as entities with unique perceptual systems that can expand our understanding of vision, cognition, and representation.
Biography:
Tom White is an artist and researcher whose work explores the intersection of human and machine perception. With over 25 years of experience in artificial intelligence and design, White studied at MIT under John Maeda, and worked alongside notable figures such as Casey Reas, Golan Levin, and Ben Fry, contributing to the foundations of creative coding tools that evolved into Processing and openFrameworks.
His art practice investigates how computational systems interpret visual information differently from humans, resulting in exhibitions worldwide that challenge viewers to consider the philosophical implications of machine perception. White's work has been recognized for revealing the surprising ways AI systems categorize and interpret the visual world, often highlighting the gaps between human intention and machine understanding.
Currently teaching at Victoria University of Wellington School of Design, White continues to blend artistic expression with cutting-edge AI research, creating work that serves not merely as technological demonstrations but as profound explorations of how machines perceive our world and what that reveals about both artificial and human intelligence.