12 February 2026

Here’s today’s concise AI + genealogy briefing for Thursday, 12 February 2026.

---

1. Major AI updates in the last 24 hours

• The Pentagon is pressing top AI companies, including OpenAI and Anthropic, to deploy their models on U.S. classified networks via the genai.mil environment, with many normal usage restrictions relaxed for defense users.[reuters]​

• OpenAI has finalized an agreement allowing its tools (including ChatGPT-style models) to be used across the Defense Department’s unclassified genai.mil network, exposing more than 3 million personnel to frontier models in routine workflows.[reuters]​

• Negotiations with Anthropic are more contentious, with the company voicing concerns about its Claude models being used for autonomous targeting or intrusive surveillance, highlighting a growing policy and ethics split among leading labs.[reuters]​

• Industry analysts are framing this week’s moves—including massive infrastructure investments and new model releases by Anthropic and OpenAI—as evidence that enterprises are rapidly normalizing “agentic” AI, where multiple agents collaborate on complex tasks.[linkedin]​

• Anthropic’s latest Claude Opus 4.6 emphasizes “agent teams,” which can divide and conquer sub‑tasks, then synthesize a unified output, a pattern likely to trickle into research, data-cleaning, and writing tools genealogists will use indirectly in coming months.[linkedin]​

• OpenAI has countered with a new coding‑oriented model to accelerate AI‑assisted software development, reflecting intensifying competition that will likely result in more rapid feature rollouts in downstream genealogy apps, plugins, and custom tools.[linkedin]​

• Defense adoption via genai.mil is reinforcing a broader shift: large organizations expect AI platforms to support multi‑model, multi‑vendor deployments, which in practice should make it easier for genealogy software and archives to integrate more than one model over time.reuters+1

---

2. Twenty+ practical AI uses for genealogists (you can try today)

Each item is framed as a concrete task pattern you can run in your preferred AI assistant (ChatGPT, Claude, Gemini, Perplexity, etc.).nwsgenealogy+3

A. Research planning and problem‑solving

1. Research question sharpening
   Paste your current brick‑wall description (e.g., a 3–5 paragraph summary) and ask the model to:

• Restate a single focused research question,

• List 5–7 specific record types and jurisdictions to prioritize, and

• Suggest a research log template tailored to that question.denyseallen.substack+1

2. Record set brainstorming for a locality
   Provide a place, time frame, and a brief profile (e.g., “Irish immigrant laborer in Chicago, 1880–1910”) and ask for a structured list of likely record types and repositories, with emphasis on lesser‑known materials like tax rolls, city engineer records, or licensing files.nwsgenealogy+1

3. Correlation matrix ideas
   Give the model a list of known facts about a person and two competing identity hypotheses, and ask it to propose a correlation table structure: columns, evidence fields, and conflict flags you can then implement in your own spreadsheet.aigenealogyinsights+1

4. Negative search planning
   Describe a failed search (which indexes checked, date ranges, name variants) and ask for a checklist of “negative search documentation” items and follow‑up repositories to contact, including suggestions for what exactly to request in an email or letter.aigenealogyinsights+1

B. Working with images and text from records

5. Handwriting‑aware transcription assistance
   Upload or paste text from a partial transcription of a difficult deed, will, or parish entry and ask the model to:

• Propose completions for bracketed gaps,

• Offer alternative readings for unclear words, and

• Flag legal or technical phrases you may want to double‑check against the image.dnapainter+1

6. Side‑by‑side paleography drills
   Create practice passages by pasting a few lines from an old record and asking the model to:

• Reproduce them in plain modern spelling,

• Explain each word’s likely 18th/19th‑century meaning, and

• Suggest 5 more sample phrases in similar handwriting for you to practice transcribing.[blog.dnapainter]​

7. Abstracting deeds and long legal instruments
   Paste a long deed or mortgage transcription and ask for a structured abstract capturing parties, relationships stated, land description, consideration, witnesses, and recording details, in a format you can paste into your research database.nwsgenealogy+1

8. Extracting people and events from parish registers
   Provide a batch of baptisms or burials (transcribed text) and ask the model to output a tidy table of names, roles (child, parent, godparent, witness), dates, places, and any explicit kinship statements that you can import into a spreadsheet.dnapainter+1

9. Name and place normalization suggestions
   Give it a list of variant spellings from your notes (surnames, villages, townlands) and ask for plausible modern standard forms, common variants, and hints for which variants to search in which jurisdictions or languages.denyseallen.substack+1

C. DNA, clusters, and relationship explanations

10. Explaining a DNA match cluster in plain language
    Paste your notes on a cluster (centimorgans, shared matches, likely MRCA candidates) and ask for a non‑technical narrative explanation suitable for sharing with a cousin: who might the common ancestor be, what other records should be checked, and what the limits of the evidence are.denyseallen.substack+1

11. Hypothesis‑drafting for unknown parentage
    Summarize your current DNA evidence and documentary context and ask the model to outline 3–4 alternative hypotheses, each with a short list of targeted documentary tests you could run to distinguish among them.aigenealogyinsights+1

12. Educational handouts for DNA classes
    Ask for a one‑page explanation of centimorgans, segment counts, and relationship ranges, written for beginners and suitable as a handout or blog sidebar, then customize and fact‑check it before use.denyseallen.substack+1

D. Timelines, maps, and context

13. Dynamic life‑event timelines
    Paste all known events for an ancestor (dates, places, sources), and ask the AI to:

• Sort them chronologically,

• Flag gaps or contradictions, and

• Suggest which gaps could be narrowed by specific records (e.g., tax lists, directories, draft registrations).nwsgenealogy+1

14. Migration route sketches
    Describe an ancestor’s probable moves (e.g., “Prussia → New York → Ohio, 1850–1875”) and ask for plausible transportation routes, ports, and waypoints, along with contemporary maps or gazetteers you should consult.aigenealogyinsights+1

15. Locality background micro‑briefings
    Ask for a 300–400 word locality profile focused on records: administrative history, boundary changes, major record‑keeping shifts, and which kinds of records survived versus were lost, then verify with archival guides.nwsgenealogy+1

E. Writing, editing, and publishing

16. Turning research notes into a narrative sketch
    Paste a messy block of research notes about one couple and ask the model to draft a one‑page narrative profile suitable for a family newsletter or blog, while preserving clear source‑attribution placeholders you will replace with your own citations.aigenealogyinsights+1

17. Source‑citation checklists (not citations themselves)
    Share one of your completed source citations (created from a style manual) and ask the AI to infer a reusable field checklist for that record type—what details you should always capture in your notes for future citations.denyseallen.substack+1

18. Plain‑language summaries of complex arguments
    When you’ve written a dense proof argument, paste it in and ask for a shorter explanation pitched at non‑specialist relatives, keeping the logic but lightening the jargon.aigenealogyinsights+1

19. Blog post idea expansion from one case study
    Feed in a short case study you’ve already written and ask for 5–10 follow‑up post ideas that reuse the same documents in different ways (e.g., methodology angle, locality spotlight, “how I found this record,” or data‑cleanup walkthrough).[aigenealogyinsights]​

20. Series planning for an ancestor or surname
    Paste a list of all posts you’ve already published about a surname or immigrant group and ask the model to propose a coherent multi‑part series outline, identifying what’s missing and where to link older posts into the new structure.denyseallen.substack+1

F. Teaching, presentations, and society work

21. Lesson outlines for beginner workshops
    Describe your audience (e.g., “library patrons who have tested at one major DNA site” or “local society members new to land records”) and ask for a 45–60 minute lesson outline with learning objectives, key examples, and simple in‑class exercises.dnapainter+1

22. Slide‑note drafting
    Provide your bullet‑point outline for an upcoming talk and ask the model to draft speaker‑note paragraphs for each slide—then edit heavily for accuracy and add your own examples.nwsgenealogy+1

23. Society newsletter content assistance
    Ask for recurring column concepts (e.g., “record type of the month,” “repository spotlight,” “member success story template”) and draft frameworks you can populate with your own local content and images.nwsgenealogy+1

24. Policy language on responsible AI use
    For a society or study group, ask the AI to propose a short policy statement on responsible, cited, and transparent AI use in research, reports, and publications, which you can refine with your board or committee.cio+1

25. FAQ drafts for your website
    Provide common questions cousins or readers ask about your methods (“Can I trust online trees?”, “Why do dates differ?”) and ask for draft FAQ answers that emphasize evidence, citations, and uncertainty, ready for you to adapt and post.denyseallen.substack+1

You can treat these as a menu: choose one active research problem or one upcoming piece of writing on your desk today, plug it into 1–2 of these patterns, and note exactly where the AI sped you up—and where traditional methods remained essential.nwsgenealogy+1