Muse Isn't a Directorial Mode.
It's Trying To Destroy History.
Ethnographic photography took images without consent for a century. The next century was spent building the rule that Muse just ignored.
Every defense of Meta's Muse Image feature eventually reaches for the same claim: photography has always taken images of people who never agreed to it. That's true. It's also not the flattering precedent people think it is.
Meta's Muse Let Anyone Remix Public Photos Without Asking
Meta launched Muse Image earlier this month through its dedicated AI unit, Meta Superintelligence Labs. One feature let users generate AI "remixes" by @-mentioning any public Instagram account, turning that account's photos into source material, with no notice and no opt in required. CAA, whose clients include Tom Hanks and Meryl Streep, told Meta directly that no one's name, image, likeness, voice, or creative work should be used by a third party, AI included, without documented consent. SAG-AFTRA raised similar concerns. Within days, Meta pulled the feature, admitting it "missed the mark."
Photography's First Era Took Images Without Ever Asking
Nineteenth and early twentieth century photography built entire archives on the premise that consent wasn't required. Anthropometric researchers like Thomas Henry Huxley and John Lamprey photographed nude subjects against standardized grids for ethnographic comparison, with no expectation of agreement from the people photographed. Orientalist photographers dressed and posed Middle Eastern and Asian subjects in costumes designed to make them look exotic and timeless for a Western market. An 1866 exhibition literally titled "The Last of the Tasmanians" treated indigenous populations as specimens to record before they vanished, not people to ask. This wasn't a fringe practice. It was the accepted working method of an entire scientific and commercial establishment, and none of it required a single signature.
The Next Century Was Spent Correcting It
After Edward Steichen's 1955 "Family of Man" exhibition tried to flatten the world's cultures into a single universal narrative, critics including Roland Barthes argued it erased history in service of a comfortable myth of shared humanity that ignored what actually separated people's circumstances, photographers spent the following decades building something deliberately different. Claudia Andujar didn't photograph the Yanomami people of northern Brazil from a distance. She lived among them for years, and her relationship with the community eventually led her to co-found an organization that spent fourteen years securing legal demarcation of Yanomami land, work that continues to this day. In South Africa, Ernest Cole photographed apartheid from inside the communities living under it, at real personal risk, having himself reclassified under apartheid law just to gain access, rather than extracting images from outside. Peter Magubane paid an even steeper price for the same choice: arrested repeatedly, held in solitary confinement for 586 days, and shot with buckshot on assignment, he kept photographing from inside the struggle rather than reporting on it from a safe distance. David Goldblatt spent over four decades building a record of daily life under segregation, and said plainly that simply living in South Africa made him complicit in what he photographed. His authority came from proximity and time, not distance and speed. None of this was incidental to the photography. It was the argument being made through the photography: that the people in the frame have a claim on how they're represented, and that claim has to be earned through relationship, not assumed through access.
The Directorial Mode Shows What Consent Actually Costs
By the 1980s, a different kind of photography made the point even more explicit. Sandy Skoglund, Laurie Simmons, and Peter Fischli and David Weiss built entire theatrical, fabricated worlds specifically to be photographed, using models, props, and constructed sets rather than found subjects. This work required contracts, paid and credited models, and licensed or built materials for every element in frame. The directorial mode wasn't photography without rules. It was some of the most heavily negotiated image making in the medium's history, precisely because everything in it had to be agreed to before the shutter opened.
This Is Not an Equivalence of Harm.
It's an Argument About Consent
I want to be precise about what I'm claiming here, because it would be easy to overreach. I am not saying that an AI remixing someone's vacation photos is equivalent to anthropometric photography of colonized peoples, or that Muse belongs in the same category of harm as apartheid-era surveillance. Those comparisons would trivialize real historical violence and I'm not making them. What I am saying is narrower: both situations turn on the same single question, whether the person in the image gets a say in how it's used. Photography's history shows that question has been fought over, lost, and slowly won across very different contexts and very different stakes. Readers may draw their own conclusions about how far that pattern extends into questions of digital extraction or platform power. That's a legitimate line of thought. It's not the argument I'm making here, and I'd rather name that distinction directly than let it sit unspoken.
Muse Didn't Return to a Baseline.
It Discarded What Was Built
Here's the correction I want to make to how I described this earlier, because the framing matters. Consent was never the starting condition of photography. It's something photographers, activists, and courts built, argument by argument, over a century and a half, from the anthropometric grid to Andujar's fourteen years of advocacy to the model release sitting in a working photographer's file cabinet. Muse doesn't return the medium to where it began. It discards what that century and a half of work produced and behaves as though none of it was ever built in the first place. Calling that a reversion to a baseline is too generous. A baseline implies neutral ground. What actually happened is closer to demolition: the work existed, it cost people real years and real risk to build, and a product team skipped past it as if it were never there.
What This Means If You Make a Living With a Camera
Every model release you've ever had signed, every usage license you've negotiated, every contract that specifies what a client can and can't do with an image after the shoot, is a small piece of the same structure Andujar spent fourteen years building at a much larger scale. That structure isn't bureaucratic overhead. It's the accumulated cost of getting photography to a point where the person in the frame has a say. Muse's opt-out default treated that structure as optional. The backlash that killed the feature in days is evidence that a lot of people, not just photographers, still understand it isn't.
Resources
Meta removes Muse Image feature after backlash, TechCrunch
CAA calls out Meta's Muse AI tool, The Hollywood Reporter
SAG-AFTRA slams Meta AI Instagram photos opt-out policy, Variety
Meta removes Muse image AI feature after backlash, Deadline
Claudia Andujar and the Yanomami Struggle, Fotomuseum Winterthur
Photographing the Yanomami, The VII Foundation
Roland Barthes's critique of "The Family of Man," background and context
Ernest Cole, photographer, Wikipedia
Peter Magubane, fearless apartheid photographer, PetaPixel
History of Photography, Module 10: From Universalism to Cultural Relativism, 1945-1975
History of Photography, Module 12: Postmodernism, Digital Photography, and the Smartphone Era