I am thirty-six years old. I was born in 1989, which means I caught the very tail end of the 1990s as a kid but have no real lived memory of the actual eighties. Everything I know about the look and feel of that era comes from old family photos, movies, and the cultural reruns that the eighties seem to keep getting on every fifteen years.
A few months ago I started seeing people post images of themselves “in” the eighties — mall portraits, yearbook photos, JCPenney-style family shots — that they had clearly generated with Nano Banana or similar tools. The images were uncanny in a specific way. They looked exactly like the photos in my parents’ albums, but the person in them was me, in clothes I had never worn, in lighting I had never been photographed under, in a year I had not yet been alive for.
I made one of myself. Then I made twenty more. Then I went down a small rabbit hole that I have not really come out of, and I want to write about what I learned because I think this is a stranger and more interesting use of image tools than the meme version makes it sound.
Why a Fake 80s Photo of Yourself Hits Differently
The first thing I noticed when I generated a convincing eighties mall portrait of myself was that I had an unexpected emotional reaction to it. Not nostalgia exactly, because I was not alive during the era. Something closer to the feeling of looking at an old family photo of someone who looks a lot like you — a younger version of a parent, an uncle in his twenties, a cousin you barely remember. The sense that this person is connected to you but also is not you.
A Nano Banana photo of yourself in an era you did not live through produces the same effect with the dial turned strange. The person in the picture is recognizably you. The era around them is recognizably not yours. The combination produces a small, slightly disorienting feeling that does not really have a name yet but that everyone I have shown these photos to seems to recognize immediately.
That feeling is, I think, why the trend exploded. It is not just a novelty. It is touching something specific about identity, memory, and the way photographs serve as evidence of having existed in a particular time.
What Makes an 80s Photo Look Like an 80s Photo
The first lesson I learned, after my early generations came back looking generic and wrong, is that a convincing period photo is built from very specific details. Most of them are not the things you would first guess.
The lighting is the biggest tell. Mall portrait studios in the eighties used a specific combination of soft frontal flash with a slightly muddy backdrop, and the resulting images had a particular flatness that no modern phone photo replicates. The color science was warmer and yellower than modern digital photos, with reds that pushed slightly into orange and skin tones that landed in a peachy register that nobody uses today.
The hair is the next biggest tell. Big hair, perms, mall bangs, side parts, mullets — these were not occasional outliers in the eighties, they were the default. A photo with modern hair on a modern face in eighties clothing reads immediately as a costume rather than a period photo. Getting the hair right is what makes the image believable.
The clothing has to be specific to the year. Early eighties is different from late eighties. Early nineties is different from late nineties. The differences are not small. A 1983 photo and a 1987 photo of the same person should not look like the same era. Generating “eighties” as a single bucket gives you generic mush. Generating “early 1985 mall portrait” gives Nano Banana enough to land in something that actually looks like a 1985 mall portrait.
And then there is the photo itself as an object — the slight curling at the corners, the faint dust speckles, the subtle color shift on the edges where the paper has yellowed over decades. Skipping these makes the image read as “person dressed in vintage clothes.” Including them makes it read as “actual photograph from the era.”
My Workflow for Getting It Right
The way I work on these now is more involved than I expected when I started. The casual workflow — upload a selfie, ask for “an 80s photo of me” — produces results that look approximately right but never really convincing.
The serious workflow starts with a clean, well-lit, neutral-expression photo of myself. The cleaner the source, the better the generated period version. Bad source photos produce bad period photos.
Then I write a very specific prompt. The year. The exact photo type — JCPenney family portrait, school yearbook headshot, disposable camera birthday party shot, vacation snapshot in a national park, family Christmas photo in a living room. The specific lighting. The hairstyle period-appropriate for the year and my apparent age in the photo. The clothing in detail. The photo medium — Kodak 400 film, Polaroid 600, instant camera with the slightly off-color flash. The signs of physical aging on the photograph itself.
I run that through Nano Banana and look at what comes back. Usually the first generation has one or two details wrong. The hair might be too modern. The shirt collar might be the wrong era. The photo quality might be too clean and digital. I adjust the prompt and generate again. Usually within three or four iterations I have a version that, if I did not know it was generated, I would believe was scanned from an old shoebox.
The Decade-by-Decade Project
Once I had the workflow down, I went a little obsessive. I used Nano Banana to generate a series of photos of myself across decades — early eighties when I would have been a small child, mid-eighties as a young kid, late eighties as a pre-teen, early nineties as a teenager, mid-nineties at a high school age, late nineties heading toward college. Each photo set in a specific scenario that would have produced an actual photo in that era.
The cumulative effect, when I lined them all up, was disorienting. It looked like the photo archive of someone whose life had taken a different turn than mine. Same person. Different timeline. A version of me that did exist in some sense and did not exist in another sense.
I sent the set to my parents. My mother sat on the phone for a long time without saying much. She said the photo of “me” at age six looked exactly like the boy I actually had been at six, except in clothing I had never owned, with a haircut I had never had, in a year before I was born. She said it felt like looking at a son who had lived a slightly different life.
I think that reaction is the actual product of this trend, more than the photos themselves.
Family Photo Series and the Memorial Use Case
The other use case I stumbled into is more emotionally serious, and I want to handle it carefully.
A few people have used Nano Banana to generate photos of family members who have died, in eras when those family members would have been young adults, doing the kinds of things they actually did. A grandfather as a twenty-five-year-old at a barbecue in 1968. A grandmother in her wedding dress in 1957. Photos that should exist but do not, because cameras were rarer in those decades and the few photos that did exist got lost across moves and generations.
I want to be clear that this is delicate territory. The generated photos are not memories. They are imagined recoveries based on present-day descriptions and old reference photos. Treating them as if they were real photos is a form of self-deception, and treating them as memorial objects without acknowledging what they are can feel hollow to the people who actually knew the person.
The use case that seems to land well is generating these photos as deliberate fictions and labeling them as such. “Here is what Grandma might have looked like at her wedding, based on the few photos we have and what we know about that year.” That framing gives the photos a place in the family memory without falsely claiming to be historical evidence.
What Nano Banana Cannot Quite Do for Period Photos
A few things to be honest about.
Specific years are easier than specific months. Asking for “August 1986” gives you August 1986 in the same generic register as 1986 generally. Nano Banana does not really know that hair styles shifted between spring and fall of 1986 the way actual style histories did.
Historical accuracy of locations is approximate. If you ask for yourself in a specific real place in a specific year — the Mall of America in 1992 — the generated background will look plausible but probably will not match what that place actually looked like in 1992. For real locations, the generated photo is interpretation, not documentation.
Cultural specificity matters. The eighties looked different in different parts of the world. American eighties is different from British eighties is different from Japanese eighties. Nano Banana defaults to American visual conventions unless you specify otherwise, and getting an authentic non-American period photo requires more deliberate prompting.
And the people in your generated photos are still you, not the friends and family who would actually have been with you in those years. Generating a believable group photo of yourself with your actual mother in 1983 requires careful character consistency work, and it still produces something that feels slightly off because the relationships between the people in the photo are imagined rather than remembered.
Why I Think This Trend Sticks Around
The mall-portrait phase of this trend will pass like most trends do. Something else will catch the internet’s attention in a few months. But I think the underlying impulse — using tools like Nano Banana to generate photos of yourself in times you did not live through — is going to keep finding new forms because it touches something real.
Photographs have always been the way we anchor identity to time. Old photos prove we existed at moments. The lack of photos from certain eras of a life can feel like an absence, like proof that those years happened in a thinner way. AI-generated period photos fill those absences with images that are not quite real but are not quite false either.
I do not think that is a problem, as long as we are honest with ourselves about what we are looking at. The photo is not a memory. It is an imagined memory, deliberately constructed. The fact that you can imagine yourself into the eighties with this clarity says something interesting about photographs as a technology, and about identity as a construction, and about the way memory and image and self have always been more entangled than we admit.
I am going to keep making these. I think a lot of people are.