Ugh I think we all have these hallucination stories when it comes to AI - I have one from yesterday, lest we think those good old days are gone! Which is the problem I have about AI as a research tool and source of information. We might recognize the errors and laugh them off, but not everyone does.
In a time when false information and fake narratives are ripping our social fabric to threads - and when more and more people are relying solely on social sources - adding more of both into the open spaces of life seems…geez… terrifying? Irresponsible? Contemptible? Wrong?
How long will it take before AI and we as a people recycle and codify those false facts into daily life as if they were true and actual?
And what will that mean for us all and the future us? 🐓🐓🌅🍁 (< chicken little 🤣)
Also @Denyse, that book seems like one you should write and put out into the world - what interesting stories you’d find!!
This is why I think it is really good for people to be aware of how these AI tools are embedded into our lives, and what they are and are not good for. I particularly like Steve Little's "For Beginner's" List of things AI is Strong for and Weaker for in Genealogy.
Strong uses:
- Summarizing. (Sometimes it over-summarizes or summarizes the wrong things.)
- Extracting. (This is really handy and has a lot of potential.)
- Generation. (It generated that whole great labor strike of 1917. This also makes it good for brainstorming and analysis, so it's a real double edged sword.)
- Translation. (Saying things a different way. Like those Shakespearean texts for my grandkids.)
- Language-related tasks. (Just as spreadsheets are very fond of numbers, Large Language Models can work magic with words. Sometimes too much magic.)
- Weaker uses:
- Research. (It’s better at processing what you’ve gathered than at gathering it, at least today. That's changing but you have to really check everything.)
- Transcription. (When you need accuracy.)
- Translation. (When subtle nuances are important.)
Thank you for including me in the list of links! I also used ChatGPT in those first weeks, foolishly asking it for books and articles on conscientious objectors in the Civil War in Pennsylvania. It promptly made up a title, author, synopsis and publisher that sounded exactly like what I wanted. After not being able to google it, I contacted Penn State Press who said “that book doesn’t exist”. But boy, I still wish it did.
Nancy - I wanted to recommend your publication on my homepage, but didn't want to do that without talking to you first. I don't see a way to message you here, thus the comment here. Please lmk if a recommendation works for you.
Ugh I think we all have these hallucination stories when it comes to AI - I have one from yesterday, lest we think those good old days are gone! Which is the problem I have about AI as a research tool and source of information. We might recognize the errors and laugh them off, but not everyone does.
In a time when false information and fake narratives are ripping our social fabric to threads - and when more and more people are relying solely on social sources - adding more of both into the open spaces of life seems…geez… terrifying? Irresponsible? Contemptible? Wrong?
How long will it take before AI and we as a people recycle and codify those false facts into daily life as if they were true and actual?
And what will that mean for us all and the future us? 🐓🐓🌅🍁 (< chicken little 🤣)
Also @Denyse, that book seems like one you should write and put out into the world - what interesting stories you’d find!!
I’ve
This is why I think it is really good for people to be aware of how these AI tools are embedded into our lives, and what they are and are not good for. I particularly like Steve Little's "For Beginner's" List of things AI is Strong for and Weaker for in Genealogy.
Strong uses:
- Summarizing. (Sometimes it over-summarizes or summarizes the wrong things.)
- Extracting. (This is really handy and has a lot of potential.)
- Generation. (It generated that whole great labor strike of 1917. This also makes it good for brainstorming and analysis, so it's a real double edged sword.)
- Translation. (Saying things a different way. Like those Shakespearean texts for my grandkids.)
- Language-related tasks. (Just as spreadsheets are very fond of numbers, Large Language Models can work magic with words. Sometimes too much magic.)
- Weaker uses:
- Research. (It’s better at processing what you’ve gathered than at gathering it, at least today. That's changing but you have to really check everything.)
- Transcription. (When you need accuracy.)
- Translation. (When subtle nuances are important.)
And I agree. @denyse. Write.
Thank you for including me in the list of links! I also used ChatGPT in those first weeks, foolishly asking it for books and articles on conscientious objectors in the Civil War in Pennsylvania. It promptly made up a title, author, synopsis and publisher that sounded exactly like what I wanted. After not being able to google it, I contacted Penn State Press who said “that book doesn’t exist”. But boy, I still wish it did.
Gosh… suddenly I’m thinking about Sarah Conner protecting her son John and stopping Starnet before Nov 30, 2022… 🤔
Nancy - I wanted to recommend your publication on my homepage, but didn't want to do that without talking to you first. I don't see a way to message you here, thus the comment here. Please lmk if a recommendation works for you.
Hm … I am not yet very Substack savvy (should be but just haven’t spent time) … let me see if I can dm you …
It could be me not figuring out! Glad we connected