|
7/19/2026 AI in Genealogy: Creating Ancestor Songs with Gemini's New Lyria 3 Music GeneratorRead NowEver since AI image generation became popular, I've wondered when AI music would become just as accessible for genealogists. I've been experimenting with AI music for a couple of years now using tools like Suno and AI Music Factory (you can read more about that at Creating Family History Songs with AI: Honoring Our Ancestor’s Through Music), usually by writing lyrics first (often with ChatGPT) and then letting the music platform bring those words to life. The results have ranged from impressive to surprisingly emotional. So when I recently noticed that Gemini now includes a Create Music feature powered by Lyria 3, I knew exactly what I wanted to test. Naturally, my official AI test subject made another appearance... My 2nd great-grandfather, Samuel "Sam" Steinman. My Test I actually discovered that this feature has apparently been available since around late February, but I simply hadn't noticed it. I spend far more time in ChatGPT and Claude than I do in Gemini, so this one slipped under my radar. For consistency, I used exactly the same source material I use for many of my AI experiments:
Test #1: No Prompt Required This may be the most beginner-friendly part of the entire feature. I simply selected the "Folk Ballad" template and uploaded Samuel's research profile. That was it. No prompt. No instructions. No songwriting. Anyone could do this. Gemini generated a song automatically. Actually...I generated two different versions, each about 30 seconds long. Version One I was pleasantly surprised. Although short, it actually included accurate information about Samuel's life and sounded pretty good. Version Two This one wasn't nearly as successful. Some basic facts were incorrect, despite using exactly the same uploaded research profile. That's one important reminder when using AI: Always fact-check the output. Wait...Why Was I Only Getting 30-Second Songs?After a little digging, I discovered something interesting. Everything I'd read suggested Lyria 3 was capable of creating songs closer to three minutes long, yet mine kept stopping after about 30 seconds. It turns out there are two different experiences:
The trick? I needed to switch my chat model from the default Flash to Pro before creating the music. Once I did that... Success. Test #2: The Three-Minute VersionNow Gemini produced a full-length song. This version was much more accurate than the second 30-second attempt. The problem wasn't accuracy. It was pacing. Samuel lived a full life. Trying to squeeze seventy years of births, deaths, moves, occupations, marriages, tragedies, and accomplishments into a three-minute song made everything feel rushed. Instead of telling a story... ...it sounded more like a timeline trying to sing. Test #3: Adding One Extra StepThis is where things became much more interesting. Instead of letting Gemini invent the lyrics, I first asked ChatGPT to write a proper folk ballad based on Samuel's research profile. Then I copied those lyrics into Gemini. I selected the same Folk Ballad music style. And let Gemini compose the music. The difference was dramatic. The lyrics flowed naturally. The story unfolded at a comfortable pace. The chorus repeated like an actual folk song. The music fit the lyrics instead of trying to summarize an entire lifetime in a few rushed verses. It sounded... Fantastic. My Overall Thoughts If you're completely new to AI... Gemini's Create Music feature is wonderful. Seriously. You don't have to know anything about prompting. You don't have to write lyrics. You don't even have to know where to begin. Choose a music style. Upload a document. Click Create. Done. That simplicity alone makes it worth trying. However... If you're willing to spend one extra minute creating better lyrics first, the results improve dramatically. For me, the winning workflow became:
Which Version Did I Like Best? Interestingly... My favorite wasn't actually the three-minute version generated entirely by Gemini. I actually preferred the original 30-second clip over Gemini's longer automatically generated song. The shorter version stayed focused and felt more polished. But the clear winner overall was: ChatGPT-generated lyrics + Gemini-generated music. That combination produced, by far, the best result. (I guess ChatGPT is still my favorite all-around AI tool!) Listen for Yourself I've uploaded all three versions to YouTube so you can compare them yourself. You'll be able to hear: I think hearing the differences side-by-side tells the story much better than words ever could. Final ThoughtsOne of my favorite parts of genealogy is finding new ways to tell our ancestors' stories. Sometimes that's through writing. Sometimes through photographs. Sometimes through animation. And... Sometimes through music. Will AI-generated songs replace carefully written family histories? Of course not. But they can add something wonderfully unexpected—a way to experience an ancestor's story rather than simply read it. And if a three-minute folk ballad inspires someone to learn more about the person behind the lyrics... I'd call that a successful experiment. More AI and Genealogy ResourcesLearn about more family history and genealogy resources under the AI in Genealogy and Genealogy Resources category and on my dedicated Genealogy Resources page.
0 Comments
Google NotebookLM's New Short Video Overview: A Genealogist's First Look
When Google NotebookLM first introduced Audio Overviews, I was impressed with how easily it could turn research into a conversational podcast. Since then, NotebookLM has continued adding new AI-powered features, and a few days ago another one quietly appeared: Short Video Overview.
Naturally, I had to try it. As many of you know, I enjoy testing new AI tools to see how they can actually help genealogists—not just because they're new, but because they might save time, improve communication, or make family history more engaging. So I spent some time experimenting. My Test
Rather than uploading random documents, I decided to use material I already knew well.
My first test used the PDF of one of my PowerPoint presentations, complete with the speaker notes: Finding Descendants in Your Family Tree NotebookLM analyzes whatever sources you add to your notebook and then offers three suggested topics it thinks would make good short videos. You can also ignore those suggestions and create your own custom topic. I generated several versions, using each option. The first attempt wasn't particularly impressive. The second was noticeably better. By the third version, NotebookLM had created a concise one-minute video that did a nice job explaining Why Genealogists Trace Descendants, one of the suggested topics. It wasn't perfect, but it was certainly good enough to share. Want to see the result? ▶ Watch: "Why Genealogists Trace Descendants" Testing with a Blog Article
Next, I wanted to see whether NotebookLM worked just as well with written content.
This time I uploaded one of my blog articles: 4 Essential Websites for Early California Genealogy Research Again, NotebookLM suggested several possible video topics. I generated several versions, trying each option, until I found one I liked. One that caught my attention was: How to Find Pre-1850 California Ancestors ▶ Watch: "How to Find Pre-1850 California Ancestors"
Testing both presentations and blog articles gave me a better idea of where this new feature shines—and where it still needs some work.
What I Like
After creating quite a few videos, here are my first impressions.
Short and Easy to Watch Each video is only about a minute long. That makes them much easier for viewers to consume than longer educational videos, especially on social media. It Finds Good Talking Points I was pleasantly surprised by many of the suggested topics NotebookLM generated automatically. Sometimes it identified the exact angle I would have chosen myself. Multiple Versions One thing I quickly learned is that you're not limited to a single result. If the first video isn't quite right, simply generate another. Each version is slightly different, and sometimes the improvement is significant. Custom Topics You're not limited to NotebookLM's suggested topics. If you already know what you'd like the video to focus on, you can enter your own custom topic and let NotebookLM build the video around that idea. Where It Still Falls Short
Like most AI tools, the results can be somewhat inconsistent.
Some videos were excellent. Others completely missed the point I wanted to emphasize. That's why I recommend generating multiple versions before deciding whether a video is worth sharing. Another thing I discovered is that the videos themselves can become fairly large files, even for these shorts. Since I wanted to add my website address to each video, I imported them into Google Vids to add a small text footer. That simple edit nearly doubled the file size. Instead of uploading those large files directly, I chose to upload them to YouTube and simply share the links. For me, that's a much more practical solution. Ideas for Genealogists
As I experimented, I found myself thinking beyond presentations and blog posts.
This feature could be useful for creating:
My Overall Thoughts
NotebookLM's Short Video Overview feature isn't perfect, but it's surprisingly useful.
Some videos will miss the mark. Others will capture your content remarkably well. The key is experimenting with different sources, trying multiple generations, and taking advantage of both the suggested and custom topics. For genealogists who already create presentations, write blog posts, or produce educational content, this feature offers another creative way to repurpose the work you've already done. I'll definitely continue experimenting with it, and I'm looking forward to seeing how Google improves it over time. If you've tried NotebookLM's new Short Video Overview feature, I'd love to hear about your experience. Did it capture your content well, or did you have to generate several versions before finding one you liked? More AI in Genealogy
Learn about more AI in genealogy updates, tips, and resources (including more NotebookLM tools), along with other family history and genealogy resources under the AI in Genealogy and Genealogy Resources categories and on my dedictaed Genealogy Resources page.
Testing Ancestry's New AI "Fill in Timeline Gaps" Feature: My First ImpressionsArtificial intelligence continues making its way into genealogy, and Ancestry has quietly introduced another AI-powered feature that caught me completely by surprise. About a week and a half ago, while I was in the middle of researching one of my ancestors, I noticed something new. As I worked through an ancestor's profile in the Facts view, I saw small "Description" buttons appearing next to some—but not all—of the timeline facts. Curious, I clicked one to see what it was. That led me to Ancestry's new BETA "Fill in Timeline Gaps" feature. Of course, I had to stop what I was doing and put it through a few tests. What Is the "Fill in Timeline Gaps" Feature? When you click one of the new Description buttons, a side panel opens displaying Ancestry's BETA Fill in Timeline Gaps tool. The feature uses AI to generate a description for a specific timeline fact. Before generating the text, you're given three options for how detailed you'd like it to be:
It's a pretty simple concept, but one that has the potential to make ancestor profiles much more informative—especially for users who don't normally add descriptions to their timeline facts. Why Fact Descriptions MatterMany Ancestry users simply enter a date and place for events like births, residences, marriages, or deaths and move on. While that's perfectly acceptable, the Description field provides an opportunity to add context that helps tell the story behind each event. For example, instead of simply recording a residence, you might note that the family was living with relatives, renting a farm, or appearing in a particular census enumeration district. A birth description might include details taken directly from a birth record, while a military fact might summarize information found in a pension file. Adding these details can make an ancestor's timeline much richer and easier to understand without having to open every attached source. That's exactly why I was interested to see how well Ancestry's new AI handled this task. My First Test The first person I tested was my fourth great-grandfather, Nehemiah Hulet, simply because I happened to be working on his profile when I discovered the feature. I selected the Birth fact and generated all three description options. This test actually impressed me. The AI stayed focused on the birth event itself.
So far, so good. My Second Test Produced Very Different Results Next, I switched to my "official tester ancestor," my second great-grandfather, Samuel Steinman, whose tree I've often used to evaluate new genealogy features. Once again, I selected the Birth fact and generated all three levels. This time, however, the results were completely different. Instead of describing Samuel's birth, the AI began summarizing his life.
That wasn't what I expected the feature to do. Testing a Residence Fact I decided to keep experimenting. I selected one of Samuel's Residence facts—one of only a couple that displayed the new Description button. This time, the results were much closer to what I had hoped for. The AI focused on the specific residence and even incorporated information from one of the three sources attached to that event. Seeing the AI reference a supporting source was encouraging because I think that's where this feature has tremendous potential. A Bit Hit-or-Miss (For Now) After trying the feature on several facts, my overall impression is that the results are currently somewhat inconsistent. Sometimes the AI stays focused exactly where it should—on the individual timeline fact. Other times it drifts into summarizing the ancestor's life rather than describing the specific event. Fortunately, this isn't particularly surprising. The feature is clearly labeled BETA, and beta features exist so developers can identify issues, improve consistency, and refine the user experience before wider release. Where I Think This Feature Could Really Shine As someone who researches professionally, I can currently write these descriptions myself much faster and with greater precision. For me, it's still quicker to manually add information to a fact description than to generate AI text and then edit it. However, experienced genealogists are not necessarily the primary audience for this feature. The majority of Ancestry users are family historians who may not realize they can—or should—add descriptions to timeline facts in the first place. For those users, this tool could become incredibly helpful. If future versions consistently:
My Overall ThoughtsI always enjoy seeing genealogy companies explore new ways to use artificial intelligence, and I appreciate that Ancestry continues experimenting with features that help users better understand and present their family history. Like many beta features, Fill in Timeline Gaps shows promise but still has a few rough edges. At the moment, I'd describe it as somewhat hit-or-miss, depending on the ancestor and timeline fact you're working with. When it stays focused on the specific event, it does a nice job of expanding the information. When it drifts into writing a mini biography for a single fact, however, it becomes much less useful. I'm optimistic that Ancestry will continue refining the feature as more people test it and provide feedback. As with any AI-generated content, remember that it should always be reviewed before accepting it. AI can be an excellent assistant, but it shouldn't replace your own analysis or careful evaluation of the records and sources supporting your research. For now, I'll continue experimenting with Fill in Timeline Gaps, and I'm interested to see how it evolves over the coming months. Have you noticed Ancestry's new AI "Fill in Timeline Gaps" feature yet? If you've tried it, I'd love to hear what you think. Did it stay focused on the individual facts, or did it wander into life summaries like some of my tests? Let me know in the comments! More Ancestry and Genealogy ResourcesLearn about more Ancestry updates, tips, tools, and features and other family history and genealogy resources under the Ancestry.com and the Genealogy Resources categories of my blog and on my dedicated Genealogy Resources page. More AI Tips, Tools, and FeaturesLearn about more AI in Genealogy tips, tools, features, and resources under the AI in Genealogy category.
MyHeritage Tribute Reel: Bringing Family Photos Together in a New Way MyHeritage continues to expand its collection of AI-powered photo tools, and in May 2026 they introduced a new feature called Tribute Reel. As someone who enjoys preserving family history and exploring new ways to share it, I decided to give it a try using photos of my mother and grandmother. The results were both fascinating and surprisingly emotional. What Is Tribute Reel? Tribute Reel lets you to combine up to 15 photographs into a single animated video. The tool animates each photo, much like MyHeritage's LiveMemory feature, then seamlessly weaves the images together into a continuous presentation. Users can:
My First Experiments I started by creating several Tribute Reels using photos of my mom. As with many AI photo tools, I quickly discovered that some images work better than others. Certain photos produced natural-looking movements, while others generated results that felt less realistic. After a few attempts, I selected the images that animated most successfully and created a version I was happy with. Next, I experimented with photos of my grandmother. Again, I tested multiple versions, swapping photographs in and out and trying different music selections to see which combinations worked best. The ability to easily recreate the reel encouraged experimentation, and I found myself refining the project several times before settling on a final version. A Genealogist's Perspective Because I personally knew both my mother and grandmother, of course, I viewed the videos with a different perspective than someone seeing them for the first time. Some of the facial expressions and movements felt familiar and believable. Others did not reflect how they actually moved or expressed themselves. This is important to remember when using any AI animation tool: the animations are generated interpretations, not authentic recordings of a person's behavior. As genealogists and family historians, we should always be careful not to confuse AI-generated content with historical reality, and mark them as so. That said, the overall effect can still be powerful. The Emotional ImpactWhat surprised me the most was the reaction from others. Several people who watched the Tribute Reel commented on how moving it was. One person said: "I don't even know you and that brought tears to my eyes. Well done!" That response reminded me that family history is ultimately about people and connections. While we spend countless hours researching records, building family trees, and analyzing DNA matches, tools like Tribute Reel offer another way to share those stories in a format that resonates emotionally with family members. For relatives who may never read a lengthy research report or family history book, a short video can sometimes create an immediate connection to the past. Final ThoughtsOf Course, Tribute Reel is not a replacement for traditional genealogy research, nor should its animations be viewed as historically accurate representations of our ancestors. However, as a storytelling tool, it offers an engaging way to showcase family photographs and create meaningful presentations for relatives. Like many AI tools, the quality of the final product depends on the photographs you choose, so don't be afraid to experiment with different images and settings. After creating reels for both my mom and grandmother, I found the feature really easy to use, fun to experiment with, and capable of producing some surprisingly touching results. For family historians looking for new ways to share treasured photographs and preserve memories, Tribute Reel may be worth exploring. Have you tried MyHeritage Tribute Reel yet? I'd love to hear about your experience and see how you've used it to bring your family history to life. More MyHeritage and Genealogy ResourcesLearn about more MyHeritage updates, tips, tools, and features and other family history and genealogy resources under the MyHeritage and the Genealogy Resources categories of my blog and on my dedicated Genealogy Resources page.
When most genealogists think about artificial intelligence (AI), they think about genealogy research. AI is often associated with transcribing documents, translating records, analyzing DNA matches, summarizing information, or helping solve brick wall problems. Although AI can certainly assist with those tasks, I use it for much more than genealogy research. Today, I spent some time creating promotional images for several of my upcoming genealogy presentations. It reminded me just how much AI has changed the way I create content for business and speaking engagements. What once required multiple programs and a considerable amount of design time can now usually be accomplished in seconds. My Journey Creating Speaking Invites with AII've been creating genealogy speaking invitations with AI ever since image generation tools first became available. Like many early adopters, I experienced the growing pains of the technology firsthand. The earliest AI-generated images often required significant editing and sometimes produced some pretty strange results. Text was frequently misspelled, designs could be inconsistent, and it usually took multiple attempts to create something usable. Over the past few years, however, the technology has improved dramatically. Each new update has brought better image quality, improved layouts, more realistic designs, and much greater control over the final result. Today, I can usually create a professional-looking speaking invitation in just a few seconds using ChatGPT's image generation capabilities. ExamplesLess than 3 minutes for the batch. AI vs. Traditional Design ToolsFor many years, I relied heavily on Canva and Adobe Express to create promotional graphics, presentation invitations, website images, and social media content. I sometimes still use both tools and find them valuable for specific projects. Canva remains useful for certain templates and recurring image series, while Adobe Express offers additional design capabilities when I need more customization. However, when I need a quick graphic for an upcoming presentation or announcement, AI has become my first choice. Instead of starting with a blank canvas and manually building a design, I can simply describe what I need and have multiple options generated within seconds. This allows me to focus more on my presentations and less on the design process. AI in Genealogy Isn't Just for ResearchOne of the biggest misconceptions about AI in genealogy is that it is only useful for research. In reality, AI can assist genealogists in many aspects of their work, whether they are professional genealogists, society volunteers, speakers, bloggers, authors, or family historians. AI has become a valuable tool not only for discovering family history but also for sharing it with others. Other Ways Genealogists Can Use AIBeyond research, genealogists can use AI to help with:
The Biggest Benefit: Saving TimeFor me, the greatest advantage of using AI for image creation is the time savings. As genealogists, we already spend countless hours researching records, analyzing evidence, developing presentations, writing reports, creating educational content, and helping others discover their family stories. The less time I spend creating graphics from scratch, the more time I can devote to serving clients, preparing presentations, writing blog posts, and conducting research. While AI won't replace creativity or expertise, it can significantly reduce the time required to complete many routine tasks. Final ThoughtsAI has become one of the most versatile tools in my genealogy toolbox. While it can certainly assist with family history research, its usefulness extends far beyond records, DNA, and brick wall problems. Whether I'm creating speaking invitations, website graphics, blog images, or educational materials, AI helps me work more efficiently and focus on the aspects of genealogy I enjoy most. The next time you think about AI in genealogy, consider looking beyond research applications. You may discover that it can help simplify many other parts of your genealogy journey as well. How are you using AI in your genealogy work? Are you using it strictly for research, or have you found creative ways to incorporate it into your genealogy projects, presentations, society work, or business? More AI in GenealogyLearn about more AI in genealogy updates, tips, and resources (including photo tools), along with other family history and genealogy resources under the AI in Genealogy and Genealogy Resources categories and on my dedictaed Genealogy Resources page.
Testing ChatGPT Images 2.0 “Research Visual” for Genealogy InfographicsOkay, so what do you do when you’re sick and playing hooky from client work? Apparently… you test AI image tools on genealogy topics. 😄 A few weeks ago, I noticed a new option inside ChatGPT Images 2.0 called “Research Visual” (complex topic infographic). Naturally, I had to try it. For my tests, I used genealogy topics and materials I already knew very well, especially my brick wall research content, so I could easily spot what worked, what didn’t, and how accurate the results actually were. I tested many examples over the last few weeks, but for this blog, I’m only showing the two brick wall examples because they showed the progression and improvements especially well, and of course I can't show them all. Test #1: Brick Wall Problem-Solving as a SystemFor the first test, I used my brick wall troubleshooting blog article as the source material, as well as my other notes. My first reaction? Very, very cool. The infographic looked polished, educational, and visually appealing. It took the structure of my article and turned it into something that felt much bigger and more layered than a normal infographic. The main sections stayed very close to my original article and were largely on point. But then ChatGPT started doing something interesting… It began adding its own extra educational sections and mini add-ons to make the infographic feel more complete and visually useful. Some of those additions were actually excellent:
Honestly? Some of the add-ons improved the infographic quite a bit. But… not everything was correct. A few sections were close, but not quite there:
The infographic kept generating resources and reading lists that looked legitimate, but several were incorrect, blended together, incomplete, or simply made up. That became a repeating pattern in the early “Research Visual” tests I tried. Test #2: Brick Wall Problem-Solving with Practical StrategiesFast forward just a couple of weeks later, and the results were noticeably different. This time:
ChatGPT simply left it out this time. And honestly? That seemed to fix the biggest problem from the earlier tests. The earlier “Research Visual” infographics kept struggling with the resources and references sections. Once those disappeared, the overall accuracy improved dramatically. The second version also felt less like a “complex topic research visual”, that it should have been, and more like a traditional infographic. So while it may have been less visually ambitious than the first test, it was far more reliable overall. And for genealogy? Accuracy matters more than fancy visuals. My Biggest Takeaways A few things really stood out to me from these tests: 1. The improvement in just a couple weeks was huge The difference between May 4 and May 23 was honestly impressive. 2. AI is getting very good at visually organizing complex genealogy concepts Especially workflows, systems, checklists, and educational-style breakdowns. 3. It still absolutely requires expert review Some of the incorrect information looked very convincing at first glance. That’s important. 4. The potential here is enormous I could easily see tools like this being useful for:
Final Thoughts Even with the flaws in the early tests, I came away really impressed with this new option. Watching the improvement happen in real time over only a couple weeks was fascinating. And honestly? I think this is just the beginning of where tools like this are heading for genealogy education and visual learning. More AI and GenealogyLearn about more AI in Genealogy updates, tips, and resources (including photo tools), along with other family history and genealogy resources under the AI in Genealogy and Genealogy Resources categories and on my dedicated Genealogy Resources page.
5/10/2026 Recreating Memories with ChatGPT Images 2.0: A Mother’s Day Journey Back to the Santa Monica MallRead Now Recreating Memories with ChatGPT Images 2.0: A Mother’s Day Journey Back to the Santa Monica MallThere are certain childhood memories that never really leave us. Not necessarily the big moments. Not birthdays or holidays. But the small, everday moments that somehow became permanent. For me, some of those memories are tied to the old outdoor Santa Monica Mall, long before it became today’s Santa Monica 3rd Street Promenade. I can still vividly remember walking there with my mom in the 1970s. Just the two of us. We would wander through the outdoor mall streets, stop at her favorite Mexican restaurant on the corner, eat chips from those classic red plastic oval baskets, and sip drinks from those swirly, textured clear plastic restaurant glasses with straws that seemed to exist in every Mexican restaurant back then. At the end of the day, we would stop by the bakery where I always got a peanut butter cookie, the kind with the crisscross fork marks pressed across the top. And somewhere in those memories was my pediatrician’s office, Dr. Levy’s Pediatrics, located right there at the mall. What fascinates me now is how specific these memories still feel. The sunlight. The walkways. The feeling of holding my mom’s hand. The colors. The textures. The atmosphere. Even stranger? I’m not entirely sure why so many of these memories are just me and my mom. Maybe my brother and sister were in school. Maybe these outings happened on doctor appointment days. Maybe it was simply our little routine together. But those memories stayed. And this Mother’s Day, I decided to try something different. Testing ChatGPT Images 2.0 for Memory RecreationA couple of weeks ago, I wrote a blog about the new ChatGPT Images 2.0 update and some of its new capabilities, including enhancement and restoration-style features. This time, however, I wanted to test something much more personal. I used one of OpenAI’s example prompts for the new image feature as inspiration, but modified it substantially to fit my own family memories and old photographs. Using actual childhood photos of me and photos of my mom as references, I prompted ChatGPT Images 2.0 to recreate scenes inspired by my memories of those outings to the Santa Monica Mall. Not exact historical recreations. Not fabricated “fake memories.” But visual interpretations inspired by real memories, real places, real feelings, and real family photographs. And honestly? The results surprised me emotionally more than technically. Seeing visual recreations of memories that had mostly existed only in my head for decades felt strangely powerful. Why This Hit Differently Than Photo RestorationAs genealogists, many of us have already experimented with AI photo restoration, enhancement, colorization, and repair tools. But this felt different. This wasn’t restoring a damaged image. This was recreating a moment. A feeling. A memory. A family story. That opens up a completely different category of possibilities for family history. New Possibilities for Family History & GenealogyAs I worked through these memory recreations, I immediately started thinking about how this could be used in genealogy and family history storytelling. Not as evidence. Not as historical fact. But as visual storytelling companions to our memories and research. Imagine being able to create memory-inspired scenes for:
But they were remembered. And storytelling has always been part of genealogy. This simply adds another visual layer to that storytelling. The Emotional Side of AI Memory RecreationOne thing I did not expect was how emotional this experiment would feel. When I saw the recreated bakery scenes with the peanut butter cookie… The Mexican restaurant scenes with the red chip basket… The pediatrician office scenes… …it genuinely felt like pieces of memory had been visually pulled back to the surface. Not perfectly. Not literally. But emotionally recognizable. That’s the important distinction. AI isn’t replacing memory here. It’s helping visualize it. There’s a difference. A New Creative Tool for GenealogistsI think this is where tools like ChatGPT Images 2.0 become especially interesting for genealogists and family historians. Not just for:
The stories behind the photographs. The memories between the photographs. The moments nobody thought to document at the time because they seemed ordinary. Until decades later when they suddenly became priceless. A Mother’s Day TributeIn the end, this little AI experiment became something more personal than I expected. A quiet tribute to my mom. A revisit to small childhood moments I hadn’t thought deeply about in years. And a reminder that sometimes the memories we carry most vividly are not the major milestones of life… …but the simple excursions. Walking through an outdoor mall. Holding your mom’s hand. Stopping for a cookie. Sitting in a restaurant booth eating chips before the meal arrives. Going to the doctor and then spending the rest of the afternoon together. Those moments mattered. And thanks to new tools like ChatGPT Images 2.0, we now have some really fascinating new ways to revisit, visualize, preserve, and share those memories with future generations. And honestly? That may be one of the most meaningful AI use cases for genealogy and family history yet. The Original Photos Behind the Memory RecreationHappy Mother’s Day to all the moms, grandmothers, and mother figures whose everyday moments became lifelong memories. More AI and GenealogyLearn about more AI in Genealogy updates, tips, and resources (including photo tools), along with other family history and genealogy resources under the AI in Genealogy and Genealogy Resources categories and on my dedicated Genealogy Resources page.
ChatGPT Images 2.0: A New Era for Visual Genealogy (and My First Tests with Photo Enhancement)On April 21, 2026, OpenAI released a major update to its image generation capabilities: ChatGPT Images 2.0. This isn’t just a minor improvement, it's a significant step forward in how artificial intelligence can create, transform, and interpret images. In OpenAI’s own words, this is: “a state-of-the-art model that can take on complex visual tasks and produce precise, immediately usable visuals… our first image model with thinking capabilities.” As someone who regularly explores how AI can be applied to genealogy, I wanted to test this new tool right away. Over the first few days, I experimented with quite a few of its features, but one stood out immediately for its practical use: Enhancing old photographs. Before diving into my results, let’s take a look at what this new tool actually offers. What Is ChatGPT Images 2.0?ChatGPT Images 2.0 is a built-in image creation and transformation tool inside ChatGPT. After clicking “Create an image,” you can either write your own prompt or select from the “Explore ideas” feature, which is a list of instant, one-click creative options, which I used. As of now, those include:
Key Capabilities (and Why They Matter for Genealogy)What makes Images 2.0 different isn’t just what it can create, but how well it understands what you’re trying to do. Looking through the examples on the OpenAI blog, here's what really stood out to me: 1. "Thinking" Capabilities This is the first image model from OpenAI with reasoning abilities. That means it doesn’t just generate visuals, it also, for the most part, understands what you're trying to do. For genealogists, this means:
2. More Realistic Images Looking through the examples on the OpenAI blog, I saw that the new model can generate pretty realistic images in a wide range of styles, including:
3. Different Styles You Can Use I also learned that you could even create:
4. Better Text in Images One of the biggest improvements: Text inside images is now actually readable and usable. This is a huge step forward for:
5. Works in Different Sizes and Layouts Whether it’s:
6. Smarter Visuals and Ideas It can now handle more complex ideas, such as:
7. Ready to Use This is where things get really interesting. This feels like something meant to be used in real, everyday situations, which I will definitely be trying next:
less time creating those visuals manually and more time focusing on research and storytelling. My First Tests: Enhancing Historical PhotosI wanted to start with something practical and useful right away: Enhancing old family photographs. Test 1: 2nd Great Grandparents Family PhotoOn the first day of its release, I tested a photograph of my 2nd great grandparents and their family (including my great grandfather). Results:
A fantastic enhancement overall but of course not a perfect preservation of every original detail. Test 2: Great Grandparents PhotoAnother strong result:
The eyes, and their expression, are everything in a photograph. Test 3: 6th Grade Friendship-Size PhotoThis was a much smaller, lower-quality image I tried. Result:
Test 4: 2nd Grade Wallet PhotoThis one was the most accurate. Result:
Overall Thoughts on Photo EnhancementThis tool is extremely promising and already useful. But it’s important to understand what it is (and isn’t): What It Does Well
Important Reminder for GenealogistsThis is where I want to be very clear: This tool enhances, but it also interprets. In other words, it is still creating parts of the image. For genealogical use:
Final ThoughtsThis is definitely a huge step up from before. We're moving from:
And this is just one of its new capabilities. In future posts, I’ll be exploring other uses I saw on the OpenAI blog:
We’re just getting started. More AI and GenealogyLearn about more AI in Genealogy updates, tips, and resources (including photo tools), along with other family history and genealogy resources under the AI in Genealogy and Genealogy Resources categories and on my dedicated Genealogy Resources page.
3/29/2026 Bringing Ancestors to Life (Again): Testing NotebookLM’s New Cinematic Video Overview FeatureRead Now When Google continues to roll out new features in NotebookLM, I pay attention—especially when those features have the potential to change how we tell family stories. Earlier this month (March 4), NotebookLM introduced a brand-new option within its Video Overviews: Cinematic Video Overview. Initially available only to Ultra users, it has now expanded to Pro users as well—which means I was able to put it to the test. And of course… I did. What Is the Cinematic Video Overview?The Cinematic Video Overview takes the standard NotebookLM video output and elevates it into something more immersive and visually engaging. Instead of a simple narrated summary, this feature:
In short: it’s designed to tell a story, not just summarize information. My First Tests (Client Work & Personal Research)I didn’t waste any time testing this out. ✔️ Client Projects My first tests were actually on client cases—and the reaction was immediate: 👉 They loved it. The earlier versions I generated felt:
Personal Test: My 2nd Great-GrandfatherNaturally, I had to test it on one of my own notebooks—one I had already built around my 2nd great-grandfather and his family. Here’s where things got… interesting. What Worked Well:
👉 AI is only as good as how it interprets the data—and sometimes it makes unexpected (or incomplete) choices. Time Investment: Not InstantOne thing to be aware of:
Instead, it feels more like: ➡️ A “set it and come back later” feature ➡️ Something you use when you want a polished storytelling output, not a quick answer First Impressions: Exciting… But EvolvingAfter multiple tests, here’s my honest take: 👍 The Pros
Genealogist Takeaway: This Is a Storytelling Tool—Not a Research ToolThis is important. The Cinematic Video Overview is not:
📄 Your research report 🎬 A shareable family story Ideas for Using This in GenealogyHere are a few practical ways you might use it:
Final ThoughtsIs it perfect? No. Is it worth experimenting with? Absolutely. What stood out to me most is this: 👉 We are moving into a space where genealogy is no longer just documented—it’s experienced. And while this feature is still evolving (and clearly changing even within weeks), it’s a glimpse into where things are heading. I’d Love to Hear From YouHave you tried the Cinematic Video Overview yet?
More AI and GenealogyLearn about more AI in Genealogy updates, tips, and resources, along with other family history and genealogy resources under the AI in Genealogy and Genealogy Resources categories and on my dedicated Genealogy Resources page.
When MyHeritage introduced Scribe AI at RootsTech 2026, it promised something genealogists have long hoped for: an AI tool capable of analyzing historical images, transcribing difficult handwriting, interpreting gravestones, and even providing insights about old family photographs. Naturally, I wanted to see how well it actually performs in real genealogy research. So instead of testing it on a few easy examples, I deliberately ran Scribe AI through a series of challenging records, including Hebrew gravestones, faded and water damaged German Bible entries, messy Spanish handwriting, a blurry newspaper screenshot, an 1848 probate letter, historical photographs, and even a degraded German parish marriage record preserved only as an old photocopy. The results were surprisingly impressive. Scribe AI is designed to analyze several types of historical images, including:
One thing that quickly became clear during my testing is that Scribe AI is not simply a transcription tool. In addition to transcribing text, the system also analyzes the image itself and provides genealogically useful insights based on the type of record being uploaded. In many cases, these insights were just as valuable as the transcription itself. The tool automatically identifies key information, historical context, and research suggestions without requiring any prompting. This is particularly useful for genealogists because the output often highlights details that might otherwise be overlooked during a quick review of a document or image. Rather than testing only one or two examples, I selected records that would challenge the system in different ways. For each example below, I am showing the Scribe AI output generated by the tool. Test 1: Jewish HeadstoneThe first test involved a Jewish headstone from a current client case. Jewish gravestones often contain valuable genealogical information such as Hebrew names, patronymics (the name of the father), and traditional Hebrew phrases or abbreviations. Scribe AI Output *Click to view Result Scribe AI successfully transcribed the Hebrew inscription and identified the structure of the information on the stone. For genealogists researching Jewish ancestry, tools like this could be especially helpful when Hebrew inscriptions contain details not found in civil records. Test 2: Blurry Jewish HeadstoneNext, I intentionally uploaded a blurry image of a Jewish headstone to test how well the system could handle poor image quality. In real genealogy research, gravestone photos are often taken under less-than-ideal conditions and may be slightly out of focus. Scribe AI Output *Click to view Result Even with the reduced image clarity, Scribe AI still produced a usable transcription of the inscription. This was an encouraging result, since many genealogy images are not perfectly photographed. Test 3: German Family Bible EntryFor the next test, I uploaded very old handwritten German entries from my 3rd great-grandfather’s family Bible. The writing was extremely challenging to read due to faded ink, messy handwriting, and water damage. Scribe AI Output *Click to view Result Scribe AI produced a strong attempt at transcribing the text. While a few names (there were 3 pages, in which only one is shown) were incorrect, the system clearly identified portions that were illegible rather than attempting to guess. This is actually a positive feature, since it prevents researchers from relying on potentially incorrect interpretations. Test 4: Spanish Handwriting on a Dark PhotocopyAnother test involved a document written in old Spanish handwriting, preserved only as a very dark photocopy from the 1960s. The background of the image is extremely dark, making it difficult to read even with the naked eye. Scribe AI Output *Click to view Result Despite the poor image quality, the transcription generated by Scribe AI was remarkably accurate and readable. Test 5: 1848 Probate LetterI also tested Scribe AI with an 1848 handwritten probate letter from one of my client cases. Older legal documents are often written in cursive styles that modern transcription tools struggle to interpret. Scribe AI Output *Click to view Result Scribe AI handled the handwriting extremely well and produced a clear transcription of the document. Test 6: Blurry Newspaper ScreenshotFor another test, I intentionally uploaded a blurry phone screenshot of a newspaper article, from another client case, to see how the system handled degraded newspaper print. Scribe AI Output *Click to view Result Despite the lower image quality, the transcription was essentially flawless. Test 7: Benchmark Death CertificateOne document I frequently use to evaluate transcription tools is my 2nd great-grandfather’s death certificate. This record contains a mixture of printed fields, handwritten entries, and medical terminology, which makes it a good test for transcription accuracy. Scribe AI Output *Click to view Result Scribe AI produced the most accurate transcription I have seen so far for this document. The only differences were the registrar’s middle initial and the word duplex in the cause of death, which was interpreted as a variation of “double.” Everything else was transcribed correctly. Test 8: Historical Family PhotographScribe AI also analyzes historical photographs, generating contextual insights about the image. I tested this feature using a photo of my 2nd great-grandmother with several of her daughters, including my great-grandmother. Scribe AI Output *Click to view Result The system generated a detailed description of the photograph and provided contextual insights about the scene and the individuals pictured. This feature could be especially helpful when analyzing unidentified or poorly documented family photos. Test 9: German Parish Marriage RecordThe most impressive result came from a German parish marriage record for my 4th great-grandparents. The image I uploaded was a very poor photocopy from the early 1980s that my father received after hiring a German researcher to extract records directly from parish registers. The page itself presents several challenges:
Scribe AI Output *Click to view Result Scribe AI transcribed nearly the entire entry correctly. The only notable differences were in my ancestor’s surname being incorrect and interpreting his father’s occupation as master mason instead of master cartwright. Considering the poor image quality and age of the record, the accuracy was impressive. Beyond Transcription: Genealogical InsightsOne of the most impressive aspects of Scribe AI is that the results go well beyond simple transcription. Depending on the type of image uploaded, the system generates different categories of analysis designed specifically for genealogical research. During my tests, the outputs included several types of insights. Gravestones For gravestone images, Scribe AI generated sections such as:
Foreign Language Documents For records written in another language, Scribe AI produced:
Documents in the User’s Language For documents written in the user interface language, the output still included:
Newspapers When analyzing newspaper articles, the system also provided:
Historical Photographs When analyzing family photographs, Scribe AI generated insights including:
Final ThoughtsAfter running Scribe AI through a wide range of real genealogical records, I came away genuinely impressed. What makes this tool particularly interesting is that it goes well beyond simple transcription. In addition to extracting text, Scribe AI analyzes the record itself and provides contextual insights, historical background, and research suggestions that are directly relevant to genealogical research. During my tests, it handled a wide variety of challenging genealogical materials including:
Of course, as with any AI tool, the results should always be reviewed carefully and verified against the original record. AI should assist genealogical research — not replace careful analysis. However, as a research assistant for extracting information and identifying clues within historical records, Scribe AI shows a great deal of promise. I will certainly continue testing it as I encounter new records in my research. For a limited time, you can try MyHeritage's Scribe AI for free. What Do You Think?AI tools are evolving rapidly in genealogy research, and Scribe AI is one of the newest tools designed specifically to help analyze historical records and images. Have you tried MyHeritage’s Scribe AI yet? If so, what types of records did you test it with? Did you find the results accurate? I would love to hear how it performs with other types of historical documents. More AI Tips, Tools, and FeaturesLearn about more AI in Genealogy tips, tools, features, and resources under the AI in Genealogy category. More MyHeritage and Genealogy ResourcesLearn about more Ancestry updates, tips, tools, and features and other family history and genealogy resources under the MyHeritage and the Genealogy Resources categories of my blog and on my dedicated Genealogy Resources page.
|
Details
Welcome to the BlogWelcome to Know Who Wears the Genes in Your Family! Here you'll find genealogy research tips, family history resources, DNA insights, technology and AI tools, genealogy news, and stories from my own research journey. Whether you're just starting your family tree, exploring your ancestry, or tackling a challenging brick wall, my goal is to help you discover, understand, and preserve your family's story. Categories
All
Top PostsBlogrollEvalogue.Life, Heart of the Family, Molly's Canopy, Climbing My Family Tree, DNA Breakthroughs, Ancestral Findings, Genealogy Tip of the Day, Family History Daily, Heart of the Family Blog Party, Genea-Musings, Best of the Genea-Blogs, Genealogy à la carte, Empty Branches on the Family Tree, Our Growing Family Tree, AI Genealogy Insights, Family Locket, Beaumont Genealogy-AI
Archives
July 2026
|

RSS Feed