Trying Out Research Like a Pro with AI Genealogy Workbook, Part 2

I’ve now moved into Chapter 2 of Diana Elder’s and Nicole Dyer’s workbook, trying to develop some new skills using AI.

My first roadblock – I used a sample Practice Prompt suggested in the workbook [involving data on direct line ancestors], but I was told I had reached my limit for today and to either upgrade to the paid version or come back in 24 hours (not tomorrow morning, but an entire day later.) That happened with both Perplexity and ChatGPT.

Gemini was my third attempt. It completed the task, but totally missed the “direct line ancestor” instruction and seems to have lied since I have no ancestor anywhere in my tree named George Sabo born in 1898. That was the first of many made up individuals.

I haven’t been impressed with Gemini when I’ve used it in the past and I’m even less impressed with it now. Gemini is a “no-go” for me from now on.

UPDATE: I waited a few days and went back to Perplexity to try out a suggested Practice Prompt with an ahnentafel PDF file attached. I was impressed with the answer. (I’m not sharing the prompts because they are proprietary to the authors.)

REPLY:

Below are up to 10 ancestors born between 1770 and 1850 in what is now Canada, as found in your ahnentafel PDF. Each entry includes the ancestor’s full name, approximate age at death, birth/marriage/death dates and places, plus names of parents if present.

[Ten ancestors listed.]

Observations

  • The above list shows a concentration of ancestors in New Brunswick, with research potential for Loyalist, Acadian, or Scottish migration stories.
  • Dates and locations are preserved as found; some further validation or deepening can be done by closely reviewing the cited ahnentafel entries.

These individuals each have ties to Canadian history during the period of 1770–1850 and should provide a rich starting point for family history research projects, especially if focusing on early communities in New Brunswick or Quebec.

I then requested three research questions based on Daniel Astle (1783-1817), one of my Canadian Loyalist family ancestors:

Here are three research questions to guide a study of Daniel Astle (1783–1817) within his historical context:

Migration and Settlement Patterns

  • How did the arrival of Loyalists and other waves of immigration after the American Revolution influence daily life and land ownership in New Brunswick and Quebec during Daniel Astle’s lifetime?

Political and Social Change

  • In what ways did the creation of New Brunswick (1784) and later administrative changes in Canada (like the division into Upper and Lower Canada) shape the identity, governance, and opportunities for British and French settlers like those in Daniel Astle’s community?

Impact of International Events

  • How did tensions and conflicts such as the aftermath of the American Revolution, border changes, and the war preparations leading up to the War of 1812 affect the security, economy, and family life of Canadian residents in the late 18th and early 19th centuries?

These questions encourage investigation into the broader historical forces shaping Daniel Astle’s world, connecting personal biography to significant social, political, and international events of the era.

These are excellent suggestions to enable me to learn more about daily life in early New Brunswick, Canada. [The Astle family was only in Quebec for a couple of years.]

The last part of Chapter 2 was focused on writing research objectives and there was an assignment to created a research objective that I could use for the remaining chapters of Research Like a Pro with AI Genealogy Workbook.

My takeaway from Chapter 2 is the importance of writing AI prompts that are filled with as many details as possible. The fewer details, the narrower the response.

On to Chapter 3!

6 thoughts on “Trying Out Research Like a Pro with AI Genealogy Workbook, Part 2”

  1. Interesting how the various AI models can be good at one thing and bad at others. In my experience, Gemini has been best at transcribing 17th century documents, but clearly, for the prompt you gave it that included a specific detail, stymied it, so it just ignored it.

    Family Historian 8 (just released) now can work with Claude and, I believe, Perplexity, but not Gemini. I’m curious to see what others report on this partnership. It seems one needs the desktop client for those models which, I imagine, cost money, so a no-go for me. But for those who are serious about integrating AI into their genealogy workflow, FH is now another option.

  2. Thank goodness there are genealogists like you who are out there exploring the various platforms and how they compare. I find in prompt writing that I’m a conversationalist – that I start with a question and a few details to start off, and then build the prompting in a series, based on question and response.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.