Renowned photographer and conceptual artist Trevor Paglen is returning to a striking source of his work: a collection of government employee portraits gathered by the Defense Advanced Research Projects Agency in the 1990s. The images were originally used to develop early facial‑recognition algorithms for the U.S. Department of Defense.
From military experiment to museum display
Paglen discovered the dataset while researching public‑domain government archives nearly a decade ago. He has since curated a series titled “It Began as a Military Experiment,” which will be featured in his 20‑year retrospective “The Most Merciful Thing in the World” at the FotoFocus Biennial in Cincinnati.
The photographs show people of varying ages, gender and skin tone against a plain sky‑blue backdrop. Paglen ran each portrait through a modern facial‑recognition algorithm, overlaying the tiny markers that indicate the points the software measures – corners of the eyes, the curve of the mouth, the edges of the nostrils – turning each person into a set of data points.
Why the work matters today
Paglen says the project forces viewers to confront how our faces have become raw material for powerful AI systems that now scan phones, airport security lines and public streets. He asks whether anyone truly consented to become part of a dataset that would later underpin surveillance tools used worldwide.
“If you’re told to do this as part of your job, do you really consent?” Paglen asks. “We retroactively become part of systems that didn’t exist at the time.”
Beyond bias – a critique of market‑driven AI
In the mid‑2010s, the conversation around facial‑recognition focused on bias – whether the technology unfairly targeted certain communities. Paglen argues that the debate misses a deeper issue: these tools are built within a capitalist market that rewards extraction of value, not fairness. He contends that once AI is embedded in market logic, discrimination becomes inevitable as companies seek profit.
“The bias conversation puts you in a position where you think it can be fixed and that equitable AI systems are possible,” he explained. “When you introduce these things into market logic, they are going to discriminate in order to extract capital from different people in different ways.”
Art, AI and the future of consent
Paglen’s work arrives at a moment when AI researchers, such as Anthropic’s Jacob Coxon, have publicly warned that advanced AI could pose existential risks. The artist sees a parallel between the awe‑inspiring yet terrifying nature of the technology and a “cosmic horror” that mirrors today’s political climate.
By exposing the origins of facial‑recognition datasets, Paglen invites the public to question how much of their own image data is already being harvested and how future applications may shape society without clear consent.
Original reporting: KTVZ (Central Oregon) — read the source article.
