Art & Datasets
ImageNet Roulette - Trevor Paglen#
Trevor Paglen’s “ImageNet Roulette” is an interactive art project that exposes the biases and stereotypes embedded within artificial intelligence systems. By allowing users to upload their photos, the AI—trained on the widely used ImageNet dataset—labels the images, often revealing unsettling and prejudiced classifications. The work highlights how AI algorithms can perpetuate harmful societal biases present in their training data.
More information: ImageNet Roulette
From ‘Apple’ to ‘Anomaly’ (Pictures and Labels) - Trevor Paglen#
In “From ‘Apple’ to ‘Anomaly’ (Pictures and Labels),” Trevor Paglen explores the ImageNet dataset, a foundational resource for object recognition in AI. By showcasing images alongside their assigned labels, Paglen reveals the subjective and often problematic ways in which AI systems categorize the world. The project encourages viewers to consider how biases are embedded in data and subsequently reflected in machine learning algorithms.
More information: From ‘Apple’ to ‘Anomaly’
The Trolls - Trevor Paglen#
In “The Trolls,” Trevor Paglen delves into the undercurrents of internet culture, examining how artificial intelligence interacts with and amplifies online behaviors like trolling. The project sheds light on the complexities of digital communication and the role AI plays in perpetuating negative social phenomena.
More information: The Trolls
It Began as a Military Experiment - Trevor Paglen#
Trevor Paglen’s “It Began as a Military Experiment” investigates the origins of surveillance technologies and their evolution from military applications to widespread civilian use.
More information: It Began as a Military Experiment
EuroNet - Constant Dullaart & Adam Harvey#
In “EuroNet” Constant Dullaart in collaboration with Adam Harvey critically examines the biases inherent in AI datasets. The project questions who controls the data that trains AI and how cultural perspectives shape machine learning.
More information: EuroNet
More Sources#
Excavating AI - Kate Crawford & Trevor Paglen#
In the essay “Excavating AI,” Kate Crawford and Trevor Paglen argue that teaching computers to interpret images is not just a technical challenge but an inherently social and political one. They suggest that despite advancements in AI, automating image interpretation will always be problematic due to the biases embedded in datasets. As AI systems increasingly influence critical decisions in society—such as hiring, education, and law enforcement—it’s vital to understand and address the politics within these technologies.
More information: Excavating AI
Trevor Paglen – Debates in AI at RISD#
In this talk at the Rhode Island School of Design (RISD), Trevor Paglen explores the intersections of art, artificial intelligence, and society. He discusses how AI technologies shape our perceptions and the ethical implications of machine learning systems. Paglen’s insights contribute to the broader conversation about the role of AI in culture and the importance of critical engagement with technology.
Watch the talk: Trevor Paglen – Debates in AI
Exhibition: AI: All Idiots#
“AI: All Idiots” is an exhibition that reveals the uncompromising and blunt language of machine learning algorithms. By creating a dataset from images downloaded from the web portfolios of Czech artists, the project highlights how AI reduces human beings to statistical data points and reinforces stereotypes. The exhibition shifts attention from individual artworks to the broader context of digital technologies, showcasing the intersection where artificial intelligence encounters and influences art.
More information: AI: All Idiots Exhibition