Ancestors and Algorithms: AI for Genealogy

Brian
  • 4.8 (31)
  • Society & Culture

Stuck on a family history brick wall? It's time to add the most powerful tool to your genealogy toolkit: Artificial Intelligence. Welcome to Ancestors and Algorithms, the definitive guide to revolutionizing your family tree research with AI. Forget the hype and confusion. This isn't just another podcast about AI; this is your hands-on, step-by-step masterclass using AI. Each week, host and researcher Brian demystifies the technology and shows you exactly how to apply AI tools to find ancestors, analyze records, and solve your toughest genealogy puzzles. We explore the incredible promise of AI while navigating its perils with an honest, practical approach. Learn to use AI as your personal research assistant—not a replacement for your own critical thinking. Join us to learn how to: Break through brick walls using AI-driven analysis and data correlation.Transcribe old, hard-to-read documents, letters, and census records in minutes.Use ChatGPT, Gemini, and other Generative AI to draft biographies, summarize findings, and organize your research.Analyze DNA matches and historical records to uncover hidden family connections.Master prompts that get you accurate results and avoid AI "hallucinations."Discover the latest AI tech and digital tools for genealogists before anyone else.Whether you're a beginner genealogist or a seasoned family historian, if you're ready to upgrade your research skills, this podcast is for you. Hit Follow now and turn AI into your ultimate secret weapon for uncovering your ancestry.

  1. 3d ago

    Ep. 58: Same Name, Two Men - Untangle a Merged Family Tree With AI

    Two men named Giovanni Farruggia registered for the World War I draft in Tampa on the same day in 1917. An online family tree fused them into one ancestor. Identity separation for genealogists: how to tell two same-name ancestors apart with ChatGPT, Claude and Gemini Notebook, using a Sicilian case from Ybor City, Tampa. This episode is for you if: You attached a record because the name matched and something feels off. Your ancestor seems to live in two houses in the same year. You suspect a FamilySearch or Ancestry tree merged two people. Your Italian, Sicilian or immigrant family reused the same names every generation. What you'll learn: The four-step identity separation method: plan the identifiers, gather every record of the name, count the people, then trace every tree fact to its owner. Why birthplace, occupation and arrival year are weak identifiers for cousins from one village. The collision test: two records that can't describe one man on one day prove two men. How a census citizenship column led to a naturalization petition that settled the case. Click-by-click setup: a ChatGPT temporary chat set to Unpersonalized, a Claude incognito chat, and Gemini Notebook (formerly NotebookLM) with one record per source. Questions this episode answers: How do I tell if two ancestors with the same name are the same person? Count before you claim. Look for collisions: two records made on the same day or in the same census that list different birth dates, addresses, wives or households. One collision proves two people. No collision proves nothing. Can AI help separate two people with the same name in a family tree? Yes, if you make it show its work. A blind Claude chat sorted ten records into two men and named its own weakest link. Gemini Notebook traced each tree fact to the man it came from, with clickable citations. Is a birthplace enough to identify an ancestor? No. Historians say Santo Stefano Quisquina supplied more than sixty percent of Tampa's Italians, so two cousins could share a village, a trade and an arrival year. How do I know if an online family tree merged two people? Trace every fact back to its source. Here the birth date and parents came from a grocer born in 1893. The wife and children came from a cigar maker born in 1895. What census column can point to naturalization records? Citizenship. The cigar maker was an alien in 1920 and naturalized by 1930, which led to his 1928 petition with his exact birth date, address, wife and children. The honest outcome: the tree's one Giovanni was two men. Rosalia's husband was identified. His father is probable and his mother is still unproved. Both AI tools made the same slip, reading "son of the head" as naming both parents. The family is a composite built from real Tampa record types. For Australian and UK listeners: the same method works with First World War service records at the National Archives of Australia, Trove, the 1939 Register for England and Wales, and FreeBMD. The Companion Guide adds advanced prompts for a side-by-side identity table and for untangling a tree merged more than once. Find it at ancestorsandai.com. Connect with Ancestors and Algorithms: 📧 Email: ancestorsandai@gmail.com 🌐 Website: https://ancestorsandai.com/ 📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/ Golden Rule Reminder: AI is your research assistant, not your researcher. Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research! New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

  2. Sep 29

    Ep. 57: Find Her Maiden Name: Proving Parents with Indirect Evidence and AI

    No record names her parents, and you can still prove them. This episode shows how, step by step, with 5 ordinary records and 3 free AI tools. Every family tree has a woman like Nancy Brockway: a first name, a married surname, and nothing before that. Our case is a composite family built from real 1850s New York record types. Nancy lived in New Berlin, Chenango County, and was born about 1816 in Otsego County. There's no marriage record and no will. There is a Hannah Pettibone, 67, in her 1850 household, and an 1847 deed that puts Nancy's name on a dead man's farm. WHAT YOU'LL LEARN - How to find a woman's maiden name with no marriage record - Why most New York marriages before 1847 left no government record - What the 1855 New York State census tells you that the 1850 census doesn't - How to search old deeds with FamilySearch Full-Text Search - How to transcribe a handwritten deed with Gemini in Google AI Studio - How to run a blind read in a Claude incognito chat, so AI can't confirm your bias - The knockout test: remove one record at a time to see what your case rests on FREQUENTLY ASKED QUESTIONS Can you prove a parent-child relationship when no record states it? Yes. Indirect evidence, several records that together leave one reasonable explanation, is fully acceptable under the Genealogical Proof Standard. Why can't I find a marriage record for my New York ancestor before 1847? According to the FamilySearch Research Wiki, no county, city, town or village government in New York recorded marriages before 1847. Statewide registration began in 1880 and 1881. How can a deed reveal a woman's maiden name? When a man died owning land and no one went to court, his heirs often settled it by deed. A married daughter may appear beside her husband with words like "in right of the said Nancy," meaning the share was hers, not his. What is a blind read with AI? You give a chatbot your records but not your theory, in a fresh incognito or temporary chat, and ask for every explanation the records allow. In this episode, Claude raised one I hadn't considered: stepmother. What is the knockout test in genealogy? Remove one record at a time and ask whether your conclusion still stands. Split your conclusion into single claims first. Claude overreached when it tested a two-part claim. Can I trust ChatGPT to tell me which genealogy records exist? Use it to brainstorm, then check every record type in the FamilySearch Research Wiki. In this episode, ChatGPT pointed to early county marriage records that don't exist for 1830s New York. THE HONEST RESULT A partial answer. The evidence best supports Nancy as a daughter of Stephen Pettibone. Whether Hannah was her mother or her stepmother is still open, and I name the record that could settle it. Australian and UK listeners: the same approach works with the NSW Historical Land Records Viewer and the National Library of Wales's wills proved before 1858. Patreon members: this episode's Companion Guide includes the full blind read and knockout workflow, plus advanced prompts for turning your results into a written proof summary. Find it, and every prompt from this episode, at ancestorsandai.com. Connect with Ancestors and Algorithms: 📧 Email: ancestorsandai@gmail.com 🌐 Website: https://ancestorsandai.com/ 📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/ Golden Rule Reminder: AI is your research assistant, not your researcher. Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research! New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

  3. Sep 22

    Ep. 56: Negative Evidence: Dating a Death in Czech Texas

    Texas did not register deaths statewide until 1903, so if your ancestor died there in the 1890s, no death certificate was ever created. This week, a county tax roll dated one anyway. Cyril Sebesta came to Fayette County, Texas from Moravia in 1873 and bought eighty acres in 1881. His headstone carries a birth year and nothing else. Three generations left a question mark in the death column. The record that finally narrowed it never mentions his death or his burial. This is a Partial Answer episode. The window closed from ten years to about twelve months, and it stops short of calling that proof. Negative evidence is suggestive, never conclusive. WHAT YOU WILL LEARN How to date a death when no death record was ever created Why searching a FamilySearch collection by name is not the same as searching the records The three gates that turn an absence into evidence Why a four-state research grid beats a three-state one How to stop AI from turning "I did not find it" into "it is not there" Free-tier workflows in Perplexity with Comet, Gemini Notebook, and Claude FREQUENTLY ASKED QUESTIONS How do I find a death date when there is no death certificate? Find a record created every year regardless of events, such as a county tax roll, a city directory, or an electoral roll, then look for the year your ancestor stops appearing in it. An annual series can narrow a death to a single year, where a ten-year census only narrows it to ten. When did Texas start registering deaths? Statewide registration of births and deaths in Texas began in 1903, and compliance was not universal at first, with rural counties lagging the cities. There is no state death certificate to order for a Texas death in the 1890s. What is negative evidence in genealogy? Negative evidence is a conclusion drawn from the absence of information where it should have appeared. It is not the same as a negative search result, which only means you looked and found nothing. Elizabeth Shown Mills calls it contextually suggestive silence. Is searching a FamilySearch collection the same as searching the records? Often not. The collection "Texas, County Tax Rolls, 1837-1910" carries a selective index covering 231 of 254 counties and only about nine years out of a seventy-year run. The rest of the images are browsable but unindexed, so a name search cannot see them. Can AI tell the difference between a missing record and one I have not searched? No. An AI tool has never seen the shelf, so it cannot separate a record that is absent from one you have not opened. That distinction has to come from you, which is why this episode uses a four-state grid instead of a three-state one. Why do Czech surnames change spelling in American records? The hacek, the small hook over the s in Sebesta, does not exist on an American pen, so clerks wrote what they heard. One family can appear as Sebesta, Shebesta, and Sebasta across three years of one record series. Identify the person by the land, not by the name. Which AI tools does this episode use, and are they free? Perplexity with Comet for record survival research, Gemini Notebook for source-grounded tabulation, and Claude for testing the argument. All three workflows run on free tiers. Gemini Notebook was renamed from NotebookLM in July 2026 and now lives at notebook.google.com. FOR AUSTRALIAN AND UK LISTENERS The same method runs on Australian electoral rolls and the digitised newspapers on Trove, and on English parish and borough rate books at the county record offices, with electoral registers on Findmypast and Scottish records on ScotlandsPeople. The Companion Guide for this episode includes the full absence grid workflow, advanced prompts, and a GPS Research Checklist built for negative evidence. It is all at ancestorsandai.com. Connect with Ancestors and Algorithms: 📧 Email: ancestorsandai@gmail.com 🌐 Website: https://ancestorsandai.com/ 📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/ Golden Rule Reminder: AI is your research assistant, not your researcher. Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research! New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

  4. Sep 15

    Ep. 55: The Blast Radius Method - Audit Your Family Tree With AI

    Most genealogists audit a family tree by hunting for blanks. The dangerous conclusions are the ones that already have a source attached. In this episode, Brian builds a fifteen-person test GEDCOM with five deliberately planted errors, then turns AI loose on it to see how many it catches. Four found. One missed. And the one it missed was the most dangerous error in the file. This is the opposite of how the genealogy world is using AI right now. Everyone is running it forward to find new ancestors. This episode runs it backward, at conclusions you already made and never revisited. You will learn the Blast Radius Method: how to rank every unsupported parent-child link in your tree by how many people collapse if that link is wrong, so you work the conclusions that matter instead of the ones that are easy. Tools used: Claude, Gemini Notebook. All on free tiers. WHAT YOU WILL LEARN How to export a GEDCOM and audit it without giving any tool write access to your treeWhy a census can never establish a birth date, and what it actually provesThe difference between a fact that is sourced and a fact that is supportedWhy unsourced facts are honest and sourced facts are the hazardHow to build a research queue sorted by consequence instead of by alphabet FREQUENTLY ASKED QUESTIONS Can AI find errors in my family tree? Yes, but only the kinds of errors you ask it to look for. In this episode's controlled test, AI caught four of five deliberately planted errors and missed an inadequate citation until asked about citations specifically. A clean report means your questions came back clean, not that your tree is clean. What is the Blast Radius Method? It is a way of prioritizing genealogy research by consequence. You rank every parent-child link in your tree by the number of descendants who depend on it, then work the links with the most people hanging off them and the least evidence behind them first. Can I use AI to analyze a GEDCOM file? Yes. A GEDCOM is plain text, so tools like Claude and Gemini Notebook can read it directly with no conversion. Export a branch rather than a whole tree, strip living people first, and treat every finding as a question to research rather than a defect to fix. Does a census prove a birth date? No. A census records an age as of the official census day, reported by an unknown household informant. The 1880 census day was June 1, 1880. A birth year derived from a census carries at least a two-year range and should never be cited as a birth date. Is a source citation the same as evidence? No. A citation is a promise that evidence exists, made by whoever typed it. A citation can be real, correct, and still inadequate for the specific fact it is attached to. A death certificate is primary evidence for the death and derivative evidence for the birth. Why is FamilySearch unable to export a GEDCOM? FamilySearch Family Tree does not offer direct GEDCOM export. To get your data into a file, you have to use one of FamilySearch's certified partner programs and export from there. THE HONEST OUTCOME The audit did not produce a clean tree. It produced a queue. And the discovery was uncomfortable: the parts of the test file that looked documented were carrying more hidden risk than the parts that were visibly blank, because a source badge is a permission slip to stop looking. Remember: AI is your research assistant, not your researcher. Companion Guide members get the full twelve-prompt audit chain, a citation adequacy table covering nineteen record classes, and a GPS Research Checklist built around evidence you already have. Connect with Ancestors and Algorithms: 📧 Email: ancestorsandai@gmail.com 🌐 Website: https://ancestorsandai.com/ 📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/ Golden Rule Reminder: AI is your research assistant, not your researcher. Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research! New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

  5. Sep 8

    Ep. 54: The Marine They Couldn't Bring Home - WWII DNA Identification & AI Genealogy Research

    A Marine killed at Tarawa in 1943 was declared non-recoverable in 1949, but DNA proved eighty-one years later he had actually been buried within days. It was the paperwork that failed him, not the men who buried him. This episode is for you if you've hit a record whose final-sounding language you never verified, if your family tree has a WWII, Korean, or Vietnam War service member whose fate was never fully confirmed, or if you want to see Gemini Notebook, Claude, and Perplexity with the Comet browser used together on one real research problem, start to finish. What you'll learn: - How to test a record's administrative language instead of accepting it at face value - How to correlate a personal service record against unit-level records to find a documented historical pattern - How modern military DNA identification actually works, and who in a family is a valid reference sample - How the Defense POW/MIA Accounting Agency and its partner History Flight are still identifying WWII remains today Frequently Asked Questions What does "non-recoverable" mean on a WWII military record? Non-recoverable is an official U.S. military case designation meaning remains could not be located or matched to a name. It does not mean no burial or recovery effort ever occurred, though families have long read it that way. How does DNA identify unidentified WWII remains? The Defense POW/MIA Accounting Agency primarily uses mitochondrial DNA, passed from a mother to all her children but carried forward only by daughters, combined with anthropological and dental analysis, to match remains against a family reference sample. Can AI tools like Claude and Gemini Notebook actually help with genealogy research? Yes. In this episode, Gemini Notebook builds a citation-backed timeline from uploaded military documents, Claude compares conflicting records to explain a contradiction, and Perplexity, using Comet, verifies the theory against current, sourced information. Why did so many Marines killed at Tarawa remain unaccounted for? Wartime construction on the tiny island of Betio disrupted cemetery markers before a full recovery could be completed in 1946. Excavations beginning around 2015 have since located dozens of graves the original recovery never found. Were Marine Corps personnel records affected by the 1973 National Personnel Records Center fire? No. That fire destroyed a majority of Army and Air Force personnel files from the affected decades, but Navy, Marine Corps, and Coast Guard records were stored separately and were largely spared. How can I find out if a missing WWII relative has been identified? Check DPAA's free "Our Missing" search tool and "Recently Accounted For" list at dpaa.mil, or ABMC's free Burials and Memorializations database at abmc.gov, which adds a bronze rosette next to a name once DPAA confirms an identification. Australian listeners: the Army's Unrecovered War Casualties unit covers more than 30,000 missing Australians from the World Wars. UK listeners: the Ministry of Defence's Joint Casualty and Compassionate Centre, the War Detectives, does the same for missing British service members. This is an Unexpected Discovery episode. The family's eighty-one-year fear that nothing had ever been recovered turned out to be wrong, though the paperwork failure behind it is real and still affects hundreds of similar cases today. A Companion Guide with twelve advanced AI prompts, a full document-comparison workflow, and a GPS Research Checklist built specifically for military identification cases is available at ancestorsandai.com. Connect with Ancestors and Algorithms: 📧 Email: ancestorsandai@gmail.com 🌐 Website: https://ancestorsandai.com/ 📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/ Golden Rule Reminder: AI is your research assistant, not your researcher. Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research! New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

  6. Sep 1

    Ep. 53: The Recording I Almost Didn't Make

    A family told the same story about a 1948 move for nearly eighty years. One recorded interview and three AI tools changed that. Effie Napier is eighty-eight, and by her family's account, the last person alive who actually remembers why they left the West Virginia coal fields for Dayton, Ohio. The story everyone told was simple: better money, a fresh start. It's the version repeated at every reunion. Then her grandson sat down with a recorder and one question built from an actual document instead of a guess. What came back wasn't the story anyone expected, including Effie, who had quietly carried the real reason alone for most of her life. This episode (using a composite case built from the kind of families and records this show sees constantly) walks through the full technique end to end: researching a relative's world with Perplexity before writing a single question, turning that research into interview questions with Claude specific enough to unlock a real memory instead of a rehearsed one, and recording, transcribing, and fact-checking the interview with Gemini Notebook, including the two-hundred-megabyte file limit that can ruin an afternoon if you don't know about it going in. This is for you if you have a living relative who remembers something your documents never will, and you want an actual method for capturing it before it's gone, not only a recorder and good intentions. FAQ What AI tools are used in this episode? Perplexity with Comet for historical research, Claude for building interview questions from documented facts, and Gemini Notebook for transcribing and analyzing the recorded interview. Is NotebookLM the same as Gemini Notebook? Yes. Google renamed NotebookLM to Gemini Notebook on July 16, 2026. It is the same product, the same notebooks, and the same underlying tool, only a new name and logo. How large can an audio file be for Gemini Notebook? The limit is two hundred megabytes per file, not a fixed number of minutes. A typical compressed recording (MP3 or M4A) can run several hours and stay under that limit. The same length recorded as an uncompressed WAV file can exceed two hundred megabytes in under thirty minutes. What questions should I ask when interviewing an elderly relative? Specific questions built from documented facts, not general prompts like "tell me about your childhood." A question tied to a real date, document, or event produces a far more detailed and useful answer. For Australian and UK listeners: the same technique applies to any internal migration story in your own family history, whether your research draws on Trove, ScotlandsPeople, or The National Archives. Honest outcome, by the show's own Genealogical Proof Standard scoring: an unexpected discovery. The facts of the timeline confirmed cleanly. What they confirmed underneath the timeline was harder and more human than anyone went looking for. The free episode covers the full technique start to finish. The Companion Guide adds twelve advanced prompts for building a repeatable oral history research practice, including how to cross-reference multiple interviews and prepare your findings for your family tree software. Find both at ancestorsandai.com. Connect with Ancestors and Algorithms: 📧 Email: ancestorsandai@gmail.com 🌐 Website: https://ancestorsandai.com/ 📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/ Golden Rule Reminder: AI is your research assistant, not your researcher. Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research! New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

  7. Aug 25

    Ep. 52: The Parish That Wouldn't Take Him Back - English Poor Law Records

    In 1836, a Yorkshire parish legally removed a laborer to a village he'd probably never lived in, over his father's old apprenticeship. This episode uses Claude, Gemini, and ChatGPT to trace an English settlement dispute through parish chest records most genealogists never open. This episode is for you if: - You have English ancestors and don't know where to start before 1837 - You've heard the term "parish chest" and want to know what's actually in one - You want to see Claude, Gemini, and ChatGPT used for real genealogy research, not generic AI tips - You're trying to understand English Poor Law, settlement, and removal records What you'll learn: - Why the GRO index cannot help you before July 1, 1837, and what to search instead - What a settlement examination and removal order actually are, and why parishes created them - How Gemini via Google AI Studio transcribes period handwriting, and which free model to pick - Why an 1836 case used a 174-year-old legal mechanism most people assume ended with Victorian workhouse reform - How to use ChatGPT for archive research strategy, not just brainstorming FREQUENTLY ASKED QUESTIONS Does the GRO index have birth records for someone born before 1837? No. Civil registration in England and Wales began on July 1, 1837, so anyone born earlier has no GRO birth record at all. Parish baptism registers are the correct record type before that date. What is a settlement examination in English genealogy? A settlement examination was a formal, sworn interview conducted by parish officials to determine which parish was legally responsible for a person if they became poor. The mechanism was created by the Settlement Act of 1662, not the older 1601 Poor Relief Act many people assume covers it. Did settlement and removal records stop after the Poor Law Amendment Act of 1834? No. The 1834 Act reorganized day-to-day poor relief into larger Poor Law Unions, but which parish was legally responsible for a specific person stayed a parish-by-parish decision until 1865. An 1836 removal order still came from an individual parish, not a Union. Can Gemini transcribe old handwritten genealogy documents for free? Yes. Gemini 3 Flash, free through Google AI Studio, performs close to human-level accuracy on legible nineteenth-century handwriting. Current testing shows the free Flash model holding up as well as, or sometimes better than, the paid Gemini Pro tier for this specific task. Can ChatGPT's free tier build a custom GPT in 2026? No. As of 2026, creating and publishing a custom GPT requires a paid ChatGPT plan. The free tier still supports unlimited text conversations, enough for research strategy and archive brainstorming. What was the Dewsbury Poor Law Union? Formed in January of 1837, the Dewsbury Poor Law Union grouped the parishes of Dewsbury, Batley, Mirfield, Thornhill, West Ardsley, and part of Birstal for administering relief under the new system. UK listeners, this one is your home turf: search the GRO's own index at gro.gov.uk, and check whether The National Archives or your county record office holds parish chest material for your ancestor's parish. Australian listeners with English-origin ancestors, Trove is worth checking for newspaper mentions of arrivals from a specific parish. This is an Unexpected Discovery, not a full breakthrough, and we say so plainly: the trail goes cold after 1836, and we don't know what happened to Amos next. Sometimes the honest answer to a four-generation mystery is smaller than the mystery itself, and that's still a real result. Want the twelve-prompt Advanced version, the exact workflow for tracing a settlement dispute through parish chest material, and a GPS Checklist for English civil and parish records? The Companion Guide has it all, plus every episode, at ancestorsandai.com. Connect with Ancestors and Algorithms: 📧 Email: ancestorsandai@gmail.com 🌐 Website: https://ancestorsandai.com/ 📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/ Golden Rule Reminder: AI is your research assistant, not your researcher. Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research! New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

  8. Aug 18

    Ep. 51: The Jewish Shtetl No Map Remembers

    A Jewish ancestor's hometown was spelled three different ways across three documents. A 1985 algorithm and three AI tools narrowed it down. This episode shows exactly how, and exactly where the method stops working. This episode is for you if you are researching Eastern European Jewish ancestors, if a hometown name on an old document does not match any map you can find, or if you want to see how Claude, Gemini, and NotebookLM actually perform on real historical documents rather than a marketing demo. You will learn: How the Daitch-Mokotoff Soundex system, built in 1985 specifically for Eastern European Jewish names, still powers the free JewishGen Communities Database todayWhy Ellis Island inspectors did not change immigrants' names, and where name changes actually happened insteadHow to use Gemini via Google AI Studio to transcribe messy historical handwriting, and how to read its uncertainty flagsHow to use Claude to weigh competing candidate towns against the evidence, and what it takes to turn a strong guess into a confirmed factHow to use NotebookLM to correlate multiple documents with citations, and why that matters when no single source closes the gapFREQUENTLY ASKED QUESTIONS What is Daitch-Mokotoff Soundex? Daitch-Mokotoff Soundex is a phonetic name-matching system built in 1985 by Gary Mokotoff and Randy Daitch specifically to handle the spelling variation common in Eastern European Jewish surnames and place names. It still powers JewishGen's free Communities Database today. Did immigration officials at Ellis Island change immigrants' names? No. Ellis Island inspectors worked from passenger manifests already prepared by the shipping line in Europe and did not write names down for the first time. When a name changed, it typically happened later, at the hands of the immigrant or a naturalization clerk. How accurate is Gemini at transcribing old handwritten documents? As of August 2026, independent testing on eighteenth and nineteenth century English-language handwriting has measured Gemini's character error rate at around one to two percent. Early Russian-language testing has been promising too, though not yet benchmarked as rigorously. What is the JewishGen Communities Database? It is a free, Soundex-searchable database that lets researchers search for a place name phonetically rather than by exact spelling, then filter results by province and by historical Jewish population, useful when a name has been transliterated inconsistently across records. Can AI tools confirm an ancestor's exact hometown? Not always, and this episode is honest about that. AI tools can narrow a list of real candidates and help weigh which one is best supported by the evidence, but confirming an exact match usually still requires a document that names the specific family in connection to that place. What AI tools are featured in this episode? Claude, for comparing and reasoning across multiple transcribed documents. Gemini via Google AI Studio, for handwriting transcription. NotebookLM, for source-grounded correlation across every document gathered, each claim cited to its source. For my Australian and UK listeners: this same wave of Eastern European Jewish emigration built major communities in Melbourne, Sydney, and London's East End. The same free Daitch-Mokotoff Soundex search works from anywhere in the world, and the National Archives of Australia and The National Archives in the UK are essential next steps once a place name is narrowed down. An honest note on the outcome: this is a Brick Wall episode. The leading candidate town is well supported but not confirmed, and the episode says so plainly rather than dressing up a strong guess as solved. For a deeper walkthrough of this exact technique, including additional advanced prompts and record-collection guidance for Eastern European Jewish research, the Companion Guide for this episode is available at ancestorsandai.com. Connect with Ancestors and Algorithms: 📧 Email: ancestorsandai@gmail.com 🌐 Website: https://ancestorsandai.com/ 📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/ Golden Rule Reminder: AI is your research assistant, not your researcher. Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research! New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

See All (63)

Trailer

4.8
out of 5
31 Ratings

About

Stuck on a family history brick wall? It's time to add the most powerful tool to your genealogy toolkit: Artificial Intelligence. Welcome to Ancestors and Algorithms, the definitive guide to revolutionizing your family tree research with AI. Forget the hype and confusion. This isn't just another podcast about AI; this is your hands-on, step-by-step masterclass using AI. Each week, host and researcher Brian demystifies the technology and shows you exactly how to apply AI tools to find ancestors, analyze records, and solve your toughest genealogy puzzles. We explore the incredible promise of AI while navigating its perils with an honest, practical approach. Learn to use AI as your personal research assistant—not a replacement for your own critical thinking. Join us to learn how to: Break through brick walls using AI-driven analysis and data correlation.Transcribe old, hard-to-read documents, letters, and census records in minutes.Use ChatGPT, Gemini, and other Generative AI to draft biographies, summarize findings, and organize your research.Analyze DNA matches and historical records to uncover hidden family connections.Master prompts that get you accurate results and avoid AI "hallucinations."Discover the latest AI tech and digital tools for genealogists before anyone else.Whether you're a beginner genealogist or a seasoned family historian, if you're ready to upgrade your research skills, this podcast is for you. Hit Follow now and turn AI into your ultimate secret weapon for uncovering your ancestry.

Information

You Might Also Like