Bringing Them Back to Life:
A Guide to AI Photo Restoration
How to sharpen, colourize, and animate old family photographs — along with the honest conversation about what these tools can do, what they cannot, and what every responsible researcher needs to know about keeping the original.
It usually happens quietly. You upload a photograph — a formal studio portrait, perhaps, from the 1890s, the kind your great-grandmother kept in her bureau — and you wait the twenty or thirty seconds while the algorithm works. And then the result appears on your screen. The face, which has been a blur of sepia tones for the past hundred and thirty years, is suddenly clear. The eyes have detail. The dress has texture. And if you have taken the extra step, the face is moving — blinking slowly, tilting slightly, offering the faintest suggestion of a smile. And something happens to you that you were not entirely expecting. You cry at your kitchen table. Or you call your sister. Or you just sit there for a while, looking at the face of someone you never met and feeling, for the first time, that you know them.
This is what AI photo restoration tools are capable of in 2026. Not just technical improvement — emotional reconnection. They are imperfect, and they come with genuine responsibilities that every researcher needs to understand. But when they work, they do something that no document or census record can: they let you look into the face of an ancestor who has been gone for a century and see them as a living person.
The first time you see a great-grandmother blink and smile in a thirty-second animation, something shifts. It is not that you forget she is gone. It is that she feels, for a moment, less completely absent.
💛 The Emotional Case: Why This Matters Beyond Technology
Before we discuss the tools themselves, it is worth dwelling on why this category of technology has had such a disproportionate emotional impact on family historians specifically. Genealogy is, at its core, an act of reaching backward — of trying to connect with people who no longer exist, through the fragmentary evidence they left behind. A census record gives you a name and an age. A deed gives you a signature. A photograph gives you a face. But until very recently, that face was static, locked in a moment, unable to look back at you.
Imagine showing your mother's face — young, before you were born, in a photograph you found in the attic — to your own grandchildren, who never knew her. The photograph is moving on the screen. She blinks. She turns slightly toward the camera. She smiles.
Your grandchildren, who have only ever known her as "Great-Grandma in the old photos," suddenly see a woman who was young once, who moved and breathed and had an expression she made when she was pleased.
This is not genealogy as data. This is genealogy as connection. And for families who have lost people recently, or who have never had video footage of an ancestor, this experience can be extraordinarily moving.
Board-certified genealogists who have tested these tools professionally describe the moment of seeing a mother's photo animated as making them cry even on a third or fourth viewing. A photograph of a mother as a little girl, in a family collection for years, can become something entirely different when colourised and animated — the child in the photograph suddenly present and specific and real in a way that black and white and stillness never quite achieved.
This emotional power is real, and it is why these tools have spread through the genealogy community so rapidly. It is also why it is important to approach them with both openness and responsibility.
🛠️ The 2026 Toolkit: What Each Tool Does
MyHeritage has built the most comprehensive suite of AI photo tools available for genealogists, and for most family historians, their platform is both the easiest starting point and the most contextually appropriate — because everything connects to your family tree rather than existing as a standalone editing experience.
🎬 Step by Step: Using Deep Nostalgia to Animate a Family Portrait
The entire process — from old photograph to animated video — takes about five minutes. Here is how to do it correctly, including the preparation steps that significantly improve the quality of the result.
🎨 Using Colourisation: A Different Kind of Connection
The emotional impact of colourisation is different from animation — quieter, perhaps, but in some ways more disorienting. We know intellectually that our ancestors' world was in colour. But years of seeing them only in black and white have created a visual association between "the past" and "grey." The moment a colourised image appears, something shifts: the past becomes less foreign. The person in the photograph stops looking like a historical figure and starts looking like a person you might pass on the street.
A formal studio portrait, ca. 1895. Woman in dark dress, pale complexion, hair pinned up. The dress appears uniformly dark. The background is a painted studio scene. Everything reads as tones of grey.
Genealogical value: name, approximate date, location (if photographer's mark survives on reverse), age, social status (formal dress indicates studio visit was a deliberate occasion).
The dress is deep burgundy-brown with subtle texture. Skin tones are warm and specific. The studio backdrop has a greenish-grey cast typical of the period. Her hair is dark auburn rather than simply "dark."
Impact: The photograph now conveys warmth, season, and the specific colour choices she made on the day she sat for the portrait — details the original could never communicate.
A colourised photograph is the AI's best estimate of the likely colours, based on training data from thousands of period-appropriate images. It is not a restoration of the "true" colours — photography in the 1890s simply did not record colour, and there is no underlying data to recover. What the AI provides is a plausible, historically-informed interpretation. It may be close to reality. It may be wrong. Treat it as a creative visualisation, not a historical document.
🏛️ Historical Integrity: The Responsibilities That Come With These Tools
Everything that follows in this section flows from one foundational principle: the original photograph is an irreplaceable historical document, and every AI tool you apply to it produces a derivative work, not a replacement.
MyHeritage acknowledges this explicitly in their own documentation: "We believe that historical accuracy and authenticity are of utmost importance. While the results of our colorization, enhancement, and animation technologies are stunning and spot-on, they are generated artificially by computer algorithms and therefore are only estimates of the true colors or details of the photos."
That acknowledgment is worth taking seriously. Here are the rules that follow from it:
- Always archive the original unedited scan. Before uploading to any AI tool, save a high-resolution copy of the original photograph in a folder marked "ORIGINALS." This file should never be overwritten, deleted, or replaced.
- Label all AI-modified files clearly. Use a consistent naming convention: 1895_Smith_Margaret_Portrait_COLOURISED.jpg, 1895_Smith_Margaret_Portrait_ANIMATED.mp4. The original should be 1895_Smith_Margaret_Portrait_ORIGINAL.jpg — and nothing should ever replace it.
- Always share both versions when sharing the photograph. When showing family the animated or colourised version, also show the original. Explain clearly that the colour and animation are AI interpretations. Future generations deserve to see the authentic historical record alongside the modern interpretation.
- Be clear when posting to social media. A colourised or animated photograph shared on social media without context can be mistaken for a genuine colour photograph or video. Always caption such images with a note: "Colourised from original black-and-white photograph using AI" or "Animated from original photograph using MyHeritage Deep Nostalgia."
- Never use AI-modified images as the primary record in your genealogy files. In RootsMagic, Legacy, or any other genealogy software, attach the original photograph to the person's record. The AI-modified version can be attached as a supplementary file, clearly labelled.
⚠️ The Deepfake Warning: Protecting Yourself and Your Family
The same AI technology that allows MyHeritage to animate your great-grandmother's portrait is related — though distinct — to the broader technology used in deepfakes: AI-generated videos that realistically put words or expressions onto a real person's face without their knowledge or consent. As family historians using these tools, there are two dimensions of this technology that you need to be aware of: the ethical responsibilities of the tools you use, and the scams targeting seniors that exploit deepfake technology.
Legitimate genealogy tools like Deep Nostalgia are not deepfakes. They apply pre-recorded movement sequences to faces in old photographs of deceased ancestors. They are transparent about what they do, they watermark every output, and they include explicit terms of service prohibiting their use to deceive. MyHeritage specifically asks users not to animate photographs of living people without their consent.
The concern is different: deepfake technology is being actively used in scams targeting seniors. Here is what to watch for:
- Scam phone calls with AI-cloned voices. Fraudsters use AI to clone a family member's voice from social media audio and call elderly relatives claiming to be in trouble and needing money. If you receive a call from a family member claiming to be in distress, hang up and call them back on a number you know to be correct.
- Video messages from "grandchildren" asking for money. AI can generate convincing video of a person's face saying things they never said. A video message is not proof that the message is genuine.
- Fake genealogy sites offering "AI photo restoration" to harvest photos of living family members. Only use photo tools from reputable, well-established genealogy platforms. Do not upload photographs of living relatives to unknown third-party AI apps — you do not know how those images will be used or stored.
- Unsolicited "restored" photographs of deceased relatives. Scam contacts sometimes approach people through genealogy forums claiming to have "found" or "restored" photographs of their ancestors, then use the interaction to request payment or personal information.
- Use only established platforms: MyHeritage, Transkribus, Google Lens, Claude, ChatGPT — not unknown apps or websites
- Read the platform's privacy policy before uploading any photograph
- Avoid uploading photographs of living family members to AI photo tools without their knowledge and consent
- Always watermark or clearly label AI-modified photographs before sharing
- If something feels wrong about an online interaction involving family photographs, trust your instinct and disengage
✦ The Gift and the Responsibility
The tools described in this guide are genuinely remarkable. They did not exist five years ago in anything approaching their current form. They will be significantly more powerful five years from now. And they do something that no previous generation of family historians could do: they let you look into a face from the 1880s and see it move.
That experience — the first time you see a great-grandparent blink, or a great-grandmother's posture shift in a LiveMemory animation — is not something that diminishes with repetition. Researchers who have been using these tools for years still describe showing new results to family members as one of the most emotionally significant things they do with their genealogy work. The technology bridges a gap that all the census records and deed books and newspaper clippings in the world cannot close: the gap between knowing someone existed and feeling, even briefly, that you knew them.
Keep the original. Label the derivative. Share both versions honestly. Use reputable tools. Stay alert to scams. And then — go find the oldest photograph in your collection, upload it, and meet whoever it contains.
Whose Face Did You
Bring Back to Life?
Tell us about your first experience with AI photo restoration — which tool you used, whose photograph you started with, and what the experience was like. Every story shared here helps another researcher understand what to expect from their own first encounter with these tools.
Share Your Story →