
Almost every secondhand shop has a box of old photos for sale. Ladies and gentlemen in their Sunday best; laborers pausing at their work; children, stiff and unnatural in a studio portrait, or candid and full of life on Christmas morning or summer vacation. Sometimes even the dead appear, the corpse surrounded by floral tributes. The one thing they all have in common is their anonymity. When the last person who knows who those people are forgets or dies, the photos go from being a vibrant archive of family history to being random scraps of ephemera destined for the landfill or for that box in the antique shop—other people’s lives for sale for a few dollars. Now, technology is beginning to offer tools to help reclaim the identity of the forgotten. At this point, the process is hit or miss, but it’s improving all the time.
Genealogy has a lot in common with detective work, and increasingly takes advantage of some of the same tools, including facial recognition software and even government databases.
The first step with photos should be digitizing them. Professional, high-resolution scanning remains the gold standard for archival materials, but that can be expensive and time consuming, and most of us are never going to commit to doing it.
A smartphone photo may not be perfect, but at least it’s easy to take and can be made instantly accessible to anyone who is interested. Cloud storage provides not only an easy way to store, organize and share family photos, it also ensures that there are copies of family treasures in a place that is safe from wildfires and other disasters.
The photo editing software that comes standard as part of the Apple Photo app is impressively powerful. It can help correct the color of film that has turned red or yellow, and it is great for lightening overexposed images, darkening underexposed photos and bringing out details.

Slides and negatives are more of a challenge but an inexpensive LED artist’s light tray and a smartphone make it relatively easy to digitize these types of media. Place the negative or slide on the light tray, and take a photo of it, then adjust the image on the phone.
Negatives need to be inverted, but that’s easy to do with the right software. Photoshop is the professional option, but free and low cost alternatives are available, including Affinity Photo, Darktable, and Gimp.
Labeling photos is just as important as digitizing them. Adding names, dates, locations, and any other identifying data to each image makes it easier to find it again, and helps with building a chronology and grouping different branches of a family together. If the people in the photo are unidentified, it’s still helpful to tag the image with keywords, even if they are based on guesswork. That mystery man with the dark eyes and the cowlick won’t care if you label him “possibly second cousin Ludwig.” If he turns out to be Great Uncle Freddy, who ran away and joined the circus (it happened) later on, all of his photos can be quickly located and relabeled.
Apple’s Photo app and Google Photos have facial recognition software built in. Once a photo is uploaded and the subject is assigned a name, the algorithm will try to identify that individual in other photos. It can often identify the same person over time, from youth to old age, and is far better at IDing babies than this researcher is. Even when the app makes a mistake it can provide information. There’s a good chance that two people who look similar enough to identify as the same person are closely related.
Genealogy services like Ancestry.com and Familysearch.org also offer their own facial recognition options, and Amazon Rekognition, AI intended for commercial image analysis, can also be used to compare photos, but a smartphone’s built-in software is highly effective, and doesn’t involve farming family photos out to third party services.
The most helpful key to photo ID is having at least one name to start with. Once you know what Great Uncle Angelo looked like chances are good that you will be able to recognize him in other photos, even without the facial recognition software, but what do you do when you’ve heard his name but never seen a photo? The answer is to talk to any relatives.
Elderly relatives may have forgotten dates and details but there is a good chance they remember faces, or can at least provide partial information. “I think that’s cousin Marie, I named my doll after her,” or “that’s my Grandfather’s garden in Indiana, he loved to grow peonies.” Even something like, “that might be my uncle Leroy, but he died before I was born” can be helpful, because if you know your relative was born in 1930, and that Uncle Leroy died before 1930, and you the clothing in the photo strongly suggests it was taken in the 1940s, you can at least eliminate Uncle Leroy from the list of possibilities, and maybe take a look at his brother Clarence.
Contemporary or younger relatives may not have known the people in the photos but they may have heard stories about them. It’s worth asking everyone you can. Sometimes other family members have copies of the same photos, and there’s always a chance that someone in that branch of the family thought to write the names on the back.
It only takes a moment to take a smartphone photo and share it. In our family it turns out that different photos of the same events were in three different locations. Having more photos helped to provide details not present in the original handful of images.
Keep trying. Sometimes one memory reactivates another. Sometimes even the most frail and forgetful relative may be having a better day and be able to tune into their memories, even if the day before they drew a complete blank.


The author’s maternal great grandmother was born in Italy in the 1860s and died in New York in the 1930s. AI was able to determine that the girl on the left in this cabinet card photograph taken at a studio in Palermo, Italy in the early 1870s is the same person as the elderly lady photographed in 1927 half a world away in her son’s garden in South Beach, New York, confirming that the older picture was a rare, childhood portrait of great grandmother. It also identified the second child as my grandmother’s favorite aunt, successfully pairing the childhood photo with one taken in the 1940s. That was truly impressive.
If the name of the person in the photo remains elusive, other details can help. Don’t just ask about names, ask about the location, the other things in the photograph, including clothing, jewelry, vehicles, furniture, and background scenery.



Google’s AI-assisted image search may be helpful for determining the date a photo was taken. It’s less good at identifying individuals: often getting the date right but misidentifying the person shown as a famous person in the same pose. Sometimes the results are hilariously wrong—Google once identified a photo taken in Upstate New York first as New Zealand and then as Norway. My Sicilian-born New York City great aunt would have been astonished to learn that Google thinks she is a famous Australian doctor. The algorithm seems to have been confused by the fact that both women were posing in 1920s cloche hats by the side of a river. That the rivers—and the women and their hats—were half a planet away didn’t trouble AI. The software gets high marks for identifying the make and model of cars, the style of clothes, and the type of photograph, but it failed more than half the time with the images we tested for this article.


AI is rapidly improving, but it is important to be aware that AI is every bit as prone to providing false or inaccurate information as that eccentric third cousin who is convinced that he is the rightful King of England, especially if it has been trained on cousin Percival’s self-published book of family history. It’s a good idea to double and triple check everything before accepting that one’s granny and aunts took a holiday in Australia in 1925, or that your Aunt Josie is really Dame Annie Jean McNamara (the Australian medical doctor and scientist) just because she’s wearing a similar hat. AI has a lot of potential as a tool for genealogists but it isn’t a substitute for independent research.
Letters and documents or all kinds, from wedding and baby announcements to that perennially verbose great aunt’s annual Christmas letters can help build not only a family tree but also help identify the people and the events in photographs.
Not everyone has family letters or papers, but there’s a good chance that a government paper trail exists for even the most elusive ancestor.
In the US, census records are an incredibly helpful place to start. All census records from 1790 to 1950 are available online (www.archives.gov). They aren’t always accurate. Some years are missing data. Often, names can be spelled wrong, and dates aren’t always accurate. Sometimes the handwriting is impossible to read, but it’s still a good place to begin, provided you have a name to start with.
Sometimes this is a process of elimination: you might not know the names of your great grandparents, but if you know where they lived and the names and ages of aunts and uncles there’s a good chance of finding the right family.
Knowing what year children were born or who lived where and when can help determine who is in a photo by process of elimination.
Those of us whose ancestors arrived in America via Ellis Island in New York have a good chance of finding those records in the archive at the Ellis Island Foundation website. However, as with census records, names can be misspelled, and data like birth years can be hard to read, easy to misread, or just plain wrong. Patience and persistence is required.
Digital photography has made it easy for researchers to document graveyards all over the country. www.findagrave.com is helpful, but so are genealogy groups on Facebook or Reddit, where volunteers may be willing to take photos of grave markers not yet documented.
Although one might think a tombstone would be the final word on a person’s birth and death dates, sometimes even here the dates are wrong. Sometimes grave markers were erected years after the person’s death, leaving room for error.
Even methodical researchers sometimes end up making mistakes. It is important to be able to let go of a theory or even a cherished family story if new research changes things. Family stories are often just that. The real story is often more compelling or horrifying than the made up one, but it can also shed light on why one’s ancestors felt the need to rewrite their narrative.There can be skeletons in anyone’s family closet: things that were hidden intentionally. Genealogy is fascinating and sometimes even thrilling, but it can also be traumatic, and can reveal old traumas.
DNA tests have enabled many to find their birth families, the names of grandparents, aunts, uncles and cousins, and connect with living relatives, but it has also rocked the world for some, leaving them grappling with their identity.
A DNA test may offer the only way forward if there are no living relatives who can help provide information, or if a person’s family is unknown. We know someone who used a popular DNA genealogy test to find the identity of her mother’s birth father, and discovered a whole new family, including a wonderful half uncle. That half uncle had been searching for his family for more than a decade before the connection was made.This was a happy outcome. Sometimes DNA revelations are not. Be prepared for the unexpected.
Many people take DNA tests for the fun of discovering their ethnic roots, something that’s still a sometimes woefully inaccurate science, and most tests remain heavily Eurocentric. One of the most useful aspects of DNA testing is the opportunity to connect with living relatives. The people who subscribe to genealogy websites and take part in DNA tests are often enthusiastic and glad to connect. This can be a good way to share photos, ask for IDs, and share family stories. Even if the newly discovered relative can’t ID a mystery photo they may be able to provide negative confirmation: “no, that’s not my father,” or provide helpful details, “I’m not sure about the people, but that looks like the family farm house in Albany.”


No matter how well-intentioned other researchers are, there is no guarantee that the research they have done is accurate. Before grafting a branch of someone else’s tree onto one’s own it’s a good idea to double check that the dates, names, and locations line up.
DNA analysis can show, with a high level of accuracy, who one’s relatives are, provided that they, too, have taken a DNA test using the same service. It’s great for finding cousins, parents, grandparents, and even long lost half siblings.
What it isn’t good at, at least not yet, is connecting the names on your family tree with the names on other people’s family trees. Anyone hoping to use it to knock down a brick wall may be out of luck. That’s because, at this point at least, the process relies on the accuracy of other people’s family trees. The fact that the guy that your third cousin identified as your mutual great great, great grandfather died twenty years before your great grandfather was born, or that the uncle on your family tree shares the same name as the uncle on someone else’s family tree and lived his whole life two thousand miles away from his doppelganger, will not stop the algorithm from trying to convince you to make a connection. Don’t believe it, not until you’ve double checked. Like everything else in 21st century genealogical research, this can be a helpful tool but it can also lead one astray—far astray. The problem isn’t the data, it’s how people use it.

Genealogy has become big business in recent years, and for many it’s a rewarding hobby or overwhelming obsession, but sometimes results are elusive. Some people may instantly turn up ancestors by the dozens, while others stare blankly at a brick wall for years. Persistence is important. New documents, books, records, and photos are being scanned all the time and added to the wealth of material available online. Someone else may have that missing puzzle piece you’ve been looking for. You might find it tomorrow, or ten years from now. And even if you never know their names, the people in family photos still tell a story. Joys and sorrows captured and preserved long after the ones who experienced that moment are dust. Genealogy offers a window into the past, and the tools that are available now have the potential to show us more than we could ever know before, but it also provides an opportunity for connection. This Thanksgiving could be an opportunity to share family photos, ask relatives about them, collect stories, and most importantly, make new memories, ones that perhaps some future genealogist may look back on with wonder and curiosity at what a weird time the early twenty-first century was.

Excellent story on photos and AI Suzanne
Brava, Suzanne!!
Beautiful writing & glorious information here,
Many thanx & I’ll be passing it along..