Thursday, March 12, 2026

12 March 2026

 

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

1. Notable AI model & tool updates

 

  • Perplexity launches Personal Computer, 

    Perplexity announced Personal Computer, a persistent AI agent that runs on a continuously running Mac mini — giving its Comet Assistant 24/7 access to your files, apps, and sessions. Personal Computer operates in a secure local environment while remaining controllable from any device, anywhere, functioning as a “persistent digital proxy of you.” Watch Personal Computer in action.

  • Anthropic upgrades Claude  for Excel and PowerPoint

    Anthropic’s Claude for Excel and Claude for PowerPoint now share full conversational context across all open files — meaning Claude can update a spreadsheet, build a model, and drop the output into a pitch deck in one continuous workflow without losing context. The update also brings Skills to both tools. Watch the updates in action.

  • ChatGPT “deep research” upgraded

    • New planning view lets you define a research plan, constrain it to specific sites or trusted apps, then watch and adjust as the run progresses, aimed at more accurate, citable long-form outputs.[help.openai]​

    • Full-screen report mode is designed for working with longer, multi-source syntheses (very relevant to locality studies and report-writing).[help.openai]​

  • Azure OpenAI releases new real-time and audio models

    • gpt-realtime-1.5 and gpt-audio-1.5 are now available through Azure, with better instruction-following, multilingual support, and low-latency interactions for voice-first applications.[learn.microsoft]​

    • Built-in PII detection can automatically flag or block sensitive personal data in outputs—useful if you’re experimenting with modern records or living-relative material.[learn.microsoft]​

  • OpenAI ecosystem trend: search-integrated reasoning

    • OpenAI’s recent o3-mini work with integrated search (late 2025) and ongoing GPT-5.x improvements show a clear push toward agents that combine web search, long context, and multi-step workflows, turning models into “desktop coworkers” rather than just chatbots.crescendo+2

  • Google’s Gemini pushes deeper into productivity tools

    • Recent Gemini upgrades across Docs, Sheets, Slides, and Drive focus on pulling structured content from your existing files and communications, auto-generating formatted docs and complex spreadsheets from natural-language prompts.[crescendo]​

    • For researchers, the key idea is AI that can reason across your stored documents (PDFs, notes, spreadsheets) instead of isolated prompts.[crescendo]​

  • Broader AI tooling landscape (high level)

    • 2026 roundups consistently list ChatGPT, Gemini, and similar assistants as the “centerpiece” of day-to-day knowledge work, with specialized video, presentation, and automation tools (e.g., Veo for video, workflow tools, slide generators) layered on top.youtube+1[synthesia]​

    • Expect more “agent-like” tools that assemble slides, videos, or briefs directly from your research packets.[synthesia]​youtube+1

2. Twenty-plus practical AI use cases for genealogists

Each of these can be done today with one of the major assistants (ChatGPT/Perplexity/Claude/Gemini) and common genealogy platforms.

Research planning, strategy, and correlation

  1. AI as a research-plan drafting partner

    • Paste a short summary of a brick-wall problem and have the AI generate a step-by-step research plan: record types, jurisdictions, time frames, and repositories to target.[denyseallen.substack]​[youtube]​

  2. Transforming scattered notes into a locality guide

    • Feed locality notes (county histories, archive descriptions, catalog excerpts) into the model and ask it to organize into a structured locality guide with sections, timelines, and key collections to consult.[youtube]​

  3. Source correlation across multiple records

    • Provide abstracts of census entries, city directories, vital records, and land deeds and ask the AI to create a correlation table, highlighting consistencies, conflicts, and possible resolutions.[denyseallen.substack]​[youtube]​

  4. Question refinement for tough problems

    • Use AI to turn a vague brick-wall description into a tightly scoped research question, complete with hypotheses and testable assertions for the FAN (Friends, Associates, Neighbors) network.[youtube]​

  5. Repository prep sheets

    • Before a visit to a courthouse or archive, paste catalog entries and policies into AI and ask for a “visit plan” listing pull-priority items, call numbers, and a checklist of lookups.[youtube]​

Working with records and languages

  1. Transcribing difficult handwriting from images or text

    • Upload or describe snippets of 18th–19th century handwriting; ask the model to produce a best-guess transcription, then refine iteratively while you verify against the image.[youtube]​

  2. First-pass translation of foreign-language records

    • Use AI to translate civil registrations, parish registers, or notarial records (e.g., Spanish, German, French, Dutch) into English, then ask for a field-by-field extraction of key genealogical data.[youtube]​

  3. Explaining archaic legal or land terms

    • Paste a land deed or probate clause and ask the model to explain unfamiliar phrases, legal customs, or land-measurement systems in plain language.[denyseallen.substack]​[youtube]​

  4. Summarizing long narrative documents

    • Ask AI to summarize multi-page obituaries, county histories, or compiled genealogies, flagging all people, places, and dates mentioned for your research log.[youtube]​

  5. Interpreting abbreviations and occupations

    • Use AI to decode abbreviations in city directories, census schedules, and occupational codes, then suggest likely industry and socio-economic context for that occupation.[denyseallen.substack]​

DNA, clustering, and inferential work

  1. Explaining DNA-match clusters in plain language

    • Export a cluster report (without personal identifiers) and ask the model to describe likely relationships, generational levels, and which cluster might tie to which ancestral couple (you still verify).youtube+1

  2. Brainstorming hypotheses for unknown ancestors using DNA clues

    • Provide anonymized segment data and match notes; have AI draft a list of possible scenarios and which records might confirm or rule out each one.[youtube]​

  3. Organizing DNA notes into a working hypothesis document

    • Paste scattered comments from your DNA notes and ask AI to reorganize into a structured hypothesis memo with sections for evidence, conflicts, and next research steps.[youtube]​

Writing, editing, and publishing

  1. Drafting ancestor biographies from research notes

    • Feed bullet-point research notes and citations for a single individual; ask AI to draft a narrative life sketch, then you revise for accuracy and voice.[aigenealogyinsights]​[youtube]​

  2. Rewriting research reports for different audiences

    • Take a technical research report and ask AI to produce: (a) a professional version retaining citations and formal language, and (b) a layperson-friendly story for a family newsletter or blog.[denyseallen.substack]​[youtube]​

  3. Creating blog post outlines from a case study

    • Paste a successful brick-wall solution and ask the assistant for a post series outline (Part 1–3), including working titles, section headings, and callouts you could fill in.aigenealogyinsights+1

  4. Improving clarity and cohesion in research logs

    • Paste a messy research log and ask AI to normalize date formats, group entries by repository, and identify gaps where a follow-up action or citation is missing.aigenealogyinsights+1

  5. Converting report prose into slides

    • Provide the text of a talk or report and ask AI to create a slide-by-slide outline with suggested headings, bullet points, and image ideas to drop into PowerPoint or Google Slides.youtube+1[synthesia]​

  6. Generating captions and descriptions for images

    • Use AI to produce clear, informative captions for historic photos, maps, or document snippets used in your blog posts, handouts, or social posts, keeping factual statements tied to your own supplied information.[synthesia]​[youtube]​

Teaching, tools, and workflow

  1. Designing lesson plans for genealogy classes

    • Ask AI to generate a 60–90 minute class plan on a topic (e.g., city directory research, land records, DNA basics) with objectives, outline, in-class exercises, and homework.[youtube]​

  2. Creating step-by-step handouts

    • Use AI to turn your internal workflow (e.g., “how I process a new record set”) into a numbered, printable checklist with screenshots you can later add.[youtube]​

  3. Building practice exercises from real records

    • Supply anonymized images or transcripts of census pages, draft cards, or vital registrations and ask AI to write student questions, answer keys, and discussion prompts around them.[youtube]​

  4. Automating FAQ responses for a genealogy blog

    • Paste a collection of reader questions and your typical answers, then work with AI to standardize and expand them into an FAQ page or chatbot-style Q&A bank.[youtube]​[synthesia]​

  5. Using AI for catalog and finding-aid navigation

    • Paste confusing archive catalog entries or finding aids and ask the model to explain what each series contains and which items are most likely to help address a particular research question.[youtube]​

  6. Storyboarding videos or short reels about research

    • Have AI turn a short written case study into a storyboard with scenes, narration points, and on-screen text suggestions for quick educational or outreach videos.[synthesia]​[youtube]​

  7. Prioritizing which record hints to review next

    • Export or manually list a set of hints from major genealogy sites (abstracting sensitive details) and ask AI to rank them by likely evidentiary value and uniqueness, so you review more impactful records first.[youtube]​[aigenealogyinsights]























at
Labels:

No comments:

Post a Comment

Subscribe to: Post Comments (Atom)