AI has opened up parts of research that you sometimes don’t see what’s right in front of you and then it will throw an idea in and you’re like wow, I never thought to look there. I’m sold on AI for transcribing and helping me with my writing and research plans.
Sometimes it can read the handwriting quite well if it’s neat cursive but old English handwriting not so good. That’s where Transkribus might be better. It’s great for correcting OCR from old newspapers but like everything ai, it still needs a human eye over it.
I used it today to help me create a PowerPoint of photos with transitions and animations. I’ve not used PP very much so rather than trying to watch a YouTube video, I asked Claude and it stepped me through the processes and guided me along the way.
Based on a suggestion from Kirsi from her use, I uploaded a page from a 1776 Quaker Monthly Meeting that I had transcribed by last week (the one referenced in my “Four AI Successes” post). It took me at least 30 minutes to finish to my satisfaction. ChatGPT produced a result in 3 seconds that, of course was as good if not more accurate than mine. And I’m using the free version of ChatGPT, not the paid. I’m a believer!
I have a couple of AI fails that I’m going to write that temper my enthusiasm. I’m probably 60/40 in terms of benefit to detriment. Some of that 40 is improper (IMO) implementation in terms of presenting declarative statements with authority to the user w/o appropriate cautions. But re: transcription, do you actually upload an image with handwriting on it and the bot returns text? I’ve never done that so I’d really be interested is you’re using something that’s affordable. Thanks for the read!
These are great examples! You gave it enough context to work with and then you tested the answers against reality (the map). I specialize in PA research and can point you in the direction of local archives that can help you learn more about these early residents.
Denyse, wow! I feel like I won the lottery connecting with an admitted PA specialist ;) Not to overstay your welcome, but just seeing your kind note I’m going to write a post about my PA ancestral roots to provide some details and sources related to my research. Thanks so much!
Looking forward to reading it! A book that may help explain the why did they move is a 1999 publication from Penn State Press called Beyond Philadelphia: The American Revolution in the Pennsylvania Hinterland. It covers the conflict inside each county at the time and is worth tracking down if you haven’t read it. I never quite grasped until I did family history that the Revolutionary War period is really 20-25 years of instability.
Thanks for the examples and responses. I’ve used AI as a brainstorming partner in my research, helping me see connections or options I might be overlooking. I’m about 50/50 on the results, but several have plaid off in new research or findings.
I think clear prompts and a willingness to question the results rather than take them as accurate are the take aways for me so far.
AI has opened up parts of research that you sometimes don’t see what’s right in front of you and then it will throw an idea in and you’re like wow, I never thought to look there. I’m sold on AI for transcribing and helping me with my writing and research plans.
Sometimes it can read the handwriting quite well if it’s neat cursive but old English handwriting not so good. That’s where Transkribus might be better. It’s great for correcting OCR from old newspapers but like everything ai, it still needs a human eye over it.
I used it today to help me create a PowerPoint of photos with transitions and animations. I’ve not used PP very much so rather than trying to watch a YouTube video, I asked Claude and it stepped me through the processes and guided me along the way.
Based on a suggestion from Kirsi from her use, I uploaded a page from a 1776 Quaker Monthly Meeting that I had transcribed by last week (the one referenced in my “Four AI Successes” post). It took me at least 30 minutes to finish to my satisfaction. ChatGPT produced a result in 3 seconds that, of course was as good if not more accurate than mine. And I’m using the free version of ChatGPT, not the paid. I’m a believer!
I have a couple of AI fails that I’m going to write that temper my enthusiasm. I’m probably 60/40 in terms of benefit to detriment. Some of that 40 is improper (IMO) implementation in terms of presenting declarative statements with authority to the user w/o appropriate cautions. But re: transcription, do you actually upload an image with handwriting on it and the bot returns text? I’ve never done that so I’d really be interested is you’re using something that’s affordable. Thanks for the read!
These are great examples! You gave it enough context to work with and then you tested the answers against reality (the map). I specialize in PA research and can point you in the direction of local archives that can help you learn more about these early residents.
Denyse, wow! I feel like I won the lottery connecting with an admitted PA specialist ;) Not to overstay your welcome, but just seeing your kind note I’m going to write a post about my PA ancestral roots to provide some details and sources related to my research. Thanks so much!
Looking forward to reading it! A book that may help explain the why did they move is a 1999 publication from Penn State Press called Beyond Philadelphia: The American Revolution in the Pennsylvania Hinterland. It covers the conflict inside each county at the time and is worth tracking down if you haven’t read it. I never quite grasped until I did family history that the Revolutionary War period is really 20-25 years of instability.
Thanks, I was looking at your website. Trying not to get overwhelmed, but as I’ve always learned, it’s all about repetition. Again.
Thanks for the examples and responses. I’ve used AI as a brainstorming partner in my research, helping me see connections or options I might be overlooking. I’m about 50/50 on the results, but several have plaid off in new research or findings.
I think clear prompts and a willingness to question the results rather than take them as accurate are the take aways for me so far.