Here’s today’s concise AI + genealogy briefing for Tuesday, 24 March 2026.

---

1. Major AI updates (last ~24 hours)

• ChatGPT has started rolling out advertisements to free and Go users in the U.S., as OpenAI looks for ways to offset high compute costs while keeping the core product widely available.[radicaldatascience.wordpress]​

• Anthropic’s Claude Cowork + Claude Code can now operate your computer by pointing and clicking, including a Dispatch feature to send tasks to your desktop from your phone, with safety controls to limit risky actions.[radicaldatascience.wordpress]​

• Anthropic models dominate Cisco’s new LLM Security Leaderboard, taking 8 of the top 10 most secure spots, with Claude Opus 4.5 ranked first for robustness against adversarial misuse.[radicaldatascience.wordpress]​

• OpenAI is in advanced talks to buy fusion-generated electricity from Helion Energy, aiming for 5 GW by 2030 and up to 50 GW by 2035, underscoring how AI demand is reshaping energy infrastructure planning.[radicaldatascience.wordpress]​

• The White House has released its first national AI policy framework, seeking to preempt state AI laws and focusing on child safety, copyright, education, and workforce training via federal standards.[radicaldatascience.wordpress]​

• Google has significantly upgraded AI Studio, adding a full‑stack “vibe coding” experience with databases, multiplayer, and persistent sessions, positioning it as a strong environment for building agentic apps and tools.[radicaldatascience.wordpress]​

• NVIDIA’s GTC announcements highlight new Nemotron 3 and Vera Rubin–class infrastructure aimed at cheaper, faster multi‑agent workloads, reinforcing NVIDIA’s central role in large‑scale AI.[radicaldatascience.wordpress]​

• OpenAI introduced GPT‑5.4 Mini and Nano as low‑latency, low‑cost “intern” models designed to act as sub‑agents under larger models for coding and multimodal tasks.[radicaldatascience.wordpress]​

• OpenAI is preparing a desktop “superapp” to unify ChatGPT’s various tools into a single interface, simplifying how users juggle chat, coding, documents, and agents.[radicaldatascience.wordpress]​

• Perplexity launched Perplexity Health, a connector‑based system for integrating personal health data from multiple sources with encrypted storage and user‑controlled access.[radicaldatascience.wordpress]​

---

2. Fast snapshot: leading engines for working genealogists

Platform	Notable for genealogists	Recent trend
ChatGPT	Strong long‑form drafting, plugins, and agent‑style workflows.clickforest+1	Moving toward ads and “superapp” desktop integration.[radicaldatascience.wordpress]​
Claude	Careful reasoning, Cowork projects, new “on‑your‑computer” automations.[radicaldatascience.wordpress]​	Strong security ranking; focus on safe agentic use.[radicaldatascience.wordpress]​
Gemini / AI Studio	Deep Google Docs/Sheets/Slides and AI Studio for apps.linkedin+1	Expanding “personal intelligence” and full‑stack building.[radicaldatascience.wordpress]​
Perplexity	Real‑time web search with citations, model switching.vezadigital+1	Health vertical launch; continued growth in search usage.[radicaldatascience.wordpress]​

A working genealogist can lean on this mix: Perplexity for current collections and tech news, Claude or ChatGPT for long proofs and reports, and Gemini/NotebookLM or AI Studio for document‑centric projects and custom tools.clickforest+2

---

3. 20+ concrete ways genealogists are using AI today

Each of these is something you could try immediately in your own research, analysis, teaching, or blogging. Draw on ChatGPT/Claude/Gemini/Perplexity interchangeably depending on context and your comfort.

1. Turning messy notes into structured research logs

   • Paste a day’s scribbles, screenshots, and URLs and have AI create a source‑by‑source log with fields like date, repository/site, collection, search terms, and outcome, ready to paste into Excel or Airtable.[last24zotero.blogspot]​

2. Summarizing long historical or legal texts

   • Feed in multi‑page deeds, chancery cases, pension files, or county histories and ask for section‑by‑section summaries, then refine into timelines or research questions.

3. Building research timelines from multiple documents

   • Paste transcribed snippets from census schedules, city directories, tax lists, and land records and have AI produce a chronological timeline for one person or couple, with each row tied back to its source.

4. Generating targeted search plans

   • Describe a brick‑wall problem, known facts, and locations, then ask AI for a prioritized search plan: record types, time frames, and repositories (e.g., “What should I check next in Oklahoma and Indian Territory for this family?”).

5. Normalizing variant name spellings at scale

   • Give AI a list of surnames from an index or cluster of autosuggest results and ask it to group likely variants (e.g., McDonald/McDonnell/MacDonnell), which you then carry into your search strategy.

6. Cleaning and standardizing place names

   • Paste a column of messy places (township abbreviations, obsolete county names, foreign diacritics) and have AI map them to standardized, time‑appropriate forms with associated jurisdictions.

7. Drafting proof arguments from structured notes

   • Once you have bullet‑pointed evidence and your reasoning, ask AI to produce a first‑draft narrative proof argument or case study, then revise for voice, add citations, and adjust for your preferred style.[last24zotero.blogspot]​

8. Line‑editing reports and articles

   • Use AI as an on‑call editor to improve clarity, transitions, and concision in research reports, journal submissions, or client letters, while you retain full control over content and conclusions.[last24zotero.blogspot]​

9. Creating multiple versions of the same family story

   • From one master narrative, generate a technical version for peers (methods, negative searches foregrounded) and a shorter, story‑first version for relatives or public talks.[last24zotero.blogspot]​

10. Explaining complex methodology in plain language

    • Ask AI to rephrase a description of FAN‑club analysis, negative evidence, or indirect evidence correlation into language suitable for beginners or for a handout.

11. Creating step‑by‑step tutorials from your workflows

    • Dictate how you use a site (for example, advanced search at FamilySearch or filtering hints at Ancestry) and have AI turn it into a numbered tutorial or checklist with clear, teachable steps.[last24zotero.blogspot]​

12. Designing short course or webinar outlines

    • Provide topic, level, and duration, then ask AI for a session outline with learning objectives, segment timings, demo ideas, and suggested homework or practice exercises.

13. Drafting blog posts from completed research

    • Paste your research log plus key findings and ask for a 800–1,200‑word blog draft, including headings and suggestions for where to place images or document snippets.[last24zotero.blogspot]​

14. Transforming blog posts into social media snippets

    • Take a finished article and have AI create short summaries, post series, or email teaser paragraphs, each pointing readers back to the full post.

15. Creating family‑friendly narrative formats

    • Ask AI to turn a factual timeline into a letter, short vignette, or “day in the life” piece, while you fact‑check and ensure every detail has documentary support.

16. Brainstorming alternative hypotheses for identity problems

    • Describe a confusing cluster of same‑name individuals in a locality and ask AI to outline several plausible identity scenarios and the specific record types that might differentiate them.

17. Assisting with foreign‑language records

    • Use AI to draft translations of civil registers, parish books, or town minutes, then have it explain key genealogical terms and typical record structures for that region and period.

18. Creating locality and migration context briefs

    • Ask AI to summarize settlement patterns, boundary changes, major economic shifts, or migration routes for a given county or region and time period, to append to your report or use as teaching background.

19. Summarizing finding aids and catalog entries

    • Paste complex archive or manuscript finding aids and have AI extract the series, dates, and specific items likely to be relevant to your target family before you place an order or plan a visit.

20. Monitoring announcements for new digital collections

    • Periodically feed AI a batch of FamilySearch, Ancestry, or state archive news posts and ask it to list only the new or expanded collections relevant to your surnames, time frame, or locations, with links and dates.[last24zotero.blogspot]​

21. Automating recurring classroom prep

    • Store a standard lesson’s key examples and prompts in an AI‑friendly format, then ask for refreshed slides outlines, quiz questions, and discussion prompts tailored to each new group.

22. Designing practice exercises from real records

    • Provide snippets from de‑identified records and have AI propose practice questions (e.g., “What further sources would you seek?” “What is the implicit relationship here?”) for students or society members.

23. Helping with citation skeletons (not finished citations)

    • Ask AI to generate basic citation patterns (elements to include, order, punctuation) for a record type and jurisdiction, then you refine into your preferred citation style and verify against the original source.

24. Assisting in repository trip planning

    • Describe a planned visit (dates, repository, research goals) and have AI propose a prioritized to‑do list: call numbers or series to check, surrounding repositories worth a side trip, and contingency tasks if specific volumes are missing.

25. Structuring multi‑generational case studies

    • When writing about a family across several generations and jurisdictions, ask AI to suggest the most reader‑friendly structure (for example, by generation, by place, or by record type) and to outline headings and subheadings you can adapt.