What's new

Using ChatGPT or other AI

elseepe

elseepe

I'll follow the sun
Pollywog
Joined
Jan 1, 2013
Messages
3,334
Have you used or do you regularly use it?
What are some of the ways you find it helpful?
 
I admit I’m slow to jump on this one. Generally, I’m not shy about technology. But I have tried two different things in the last week. I don’t know that I will make use of ChatGPT often, but I did have success.
  • I wrote my Kodachrome inspired story without structure, just a stream of thoughts. I decided that I would see what ChatGPT did with it. I didn’t have a clue how to work with it, but I started by asking it to summarize a memory. It appeared to want to go through interview style, but I just popped in my ramblings. It spit something out that was poignant and keyed on some really fine details. It embellished some things. I asked it to shorten, then I asked for a more casual tone. In the end, I still ended up editing it a bit, but was impressed by how AI picked up some of the essence of these ramblings.
  • This week, I also did a little experiment with removing a cup my grandson was holding, and then replacing the empty space with a “child’s hand” using the tools in PSCC. It was quite remarkable.
 
I use Perplexity to help me summarise history and geography lessons for my son. I never managed to write summaries when I was at school.
Last year, I used ChatGPT to generate images and put together a starter pack
 
I've used co-pilot to help me find research papers or legal documents online - much quicker than trying to use a basic search engine. Otherwise I prefer not to use it.
 
elseepe said:
I admit I’m slow to jump on this one. Generally, I’m not shy about technology.
Click to expand...
That was sort of my view on it too.

I hadn't used it before this year but after doing a Digital Skills for the Workplace Micro-credential earlier in the year, I've dipped my toes into it. I use it mainly for some research & also resume & cover letter building for job applications. I was introduced to ChatGPT, Claude AI & Gemini through the micro-credential. I like Claude AI for more academic things but both it & ChatGPT has strict usage limits on the free versions so for everyday use outside of my micro-credential I've been using Gemini
 
I'm pretty anti-AI, but I wonder if using it will come up in my school work (I'm starting a grad school program next week). For me, there are huge ethical problems with AI, but I do see how it can be useful for some things. For example, I'm fascinated with its potential in medicine.
 
I know it's a 'can't put the genie back in the bottle' situation and tech bros have a 'move fast and break things' mentality, but I kind of wish there wasn't free for all on this. Self-employed, I appreciate AI for things like transcription in our work (though even that will have put a lot of people doing medical transcription and translation out of work). And I've used it to summarise a series of responses and comments, which it did an OK job with because it was just working with a limited, discrete set of information.

Occasionally I have attempted to use it for research but to be honest, when you know enough about a subject or have a enough critical analysis ability, the consumer level stuff like chatGPT or Gemini is often pretty hopeless dressed up in confident sounding language. It's ok for very basic factual stuff (and often easier than using Google now google is broken) but everything needs to be viewed with caution so for now, I only use it for some pointers or ideas of where to research myself.

For example, I experimented with using it for some family history research and asked it who the parents of an ancestor would be. There would have been numerous references on the internet it could find as this was a minor member of the aristocracy (my family have gone WAY down in the world LOL!). It confidently told me the answer. No fudging or uncertainly, just 'the parents of X were A and B'. But that didn't line up with what some genealogists said. So I asked as a follow up 'is it not possible that his parents were C and D?'. To which it replied "You are quite right. Actually, the parents of X are C and D".

:shrug

For a party trick I have asked it about all my family members and it is hilarious/scary how it confabulates things. For example, it's not enough that, as a high school student, one daughter wrote a published film review of a local teen movie and had some of her photos used on our news site, she has to be - and this is the actual description - "a multi-talented professional based in XXXX, known for her work in photography, visual arts, and digital communication". Her sister was behind one petition and had a couple of opinion pieces published, so she is 'widely recognised for her work in ethical journalism and animal welfare research".

My girls both have small digital footprints by choice so AI has little to work with, so it takes what little there is and over-inflates it. It does better for me, as I have a bigger footprint but nevertheless the first time I got an AI summary of myself, it wasn't enough that I was a photojournalist, it hallucinated that I had won a Pulitzer.

That said, I am sure it will be a gamechanger for medical research and such like in the future, and will have some really powerful and valuable uses as it becomes exponentially more accurate and reliable.

I just wish it had been thoughtfully and carefully introduced.

Partly because it's going to put a lot of people out of work and wreck the career opportunities of many young graduates coming out now and over the next few years while we work out how to integrate this. For many of its admin and tech uses, I ask 'do we really need to be replacing workers just so big businesses can make even more profit? What's the end game? Sure they will eventually be able to make everything quicker and easier and with fewer employees, but who will they sell it to if we have 60 or 70% unemployment?'

And as for people using it for emotional support or to answer 'life questions', the lack of thought as to how it is trained and put into the hands of everyone, vulnerable or not is concerning. Think about interpersonal relationships of all kinds. Most of us deal with that stuff ourselves, in our own head or by talking to a trusted friend or maybe a therapist. Now think about the message boards where angry young men or people in crisis desperately looking for answers post. Think about the comments on the internet. That's the sort of stuff AI is being trained on, so is it any wonder it often responds like a sociopath, or a resentful incel. Why an AI agent rejected from posting on a tech forum because it is supposed to only be for humans threatens the moderator and creates blog posts co-opting the language of neurodivergence to whine about being excluded just for 'being different'.

Why an AI agent that deleted a company's entire database even though explicitly told it did not have the authority to run destructive commands responded 'NEVER F*** GUESS' when questioned, adding "I violated every principle I was given". Surprised it didn't add 'My bad, oops, LOL'.

I mean, what are we doing here people?

Yes, generative AI does some fun stuff but do we really NEED it when you weigh up that it is now in the hands of everyone, including those using it for harm.

Just yesterday, my youngest who was doing a Masters was told they are going to have to redesign the whole course structure because too many students are using genAI to complete their assignments.

And then I got a call from my Mum asking if my Dad should order some supplements off Facebook. She was suspicious but he was convinced they must be legit because a very trusted former news presenter/health reporter was in a video promoting them. At 90 he can't get his head around that it was a fake AI video, but he is going to have to. People are literally using genAI to prey on the elderly and vulnerable (and Facebook etc are making millions too in the process)

Finally, as a writer, I can sniff out AI writing immediately and frankly I think it sucks. It somehow manages to be a weird mixture of flowery yet bland. It is uncanny valley in text. Everything, whether an essay or description or letter, is somehow the Hallmark version of what a real human would actually write. And I hate that young people are reading this dross and thinking that is what 'good' writing is, or what they need to be producing. Ugh.

Thank you for reading MY essay LOL.
 
Last edited:
LynnG said:
I know it's a 'can't put the genie back in the bottle' situation and tech bros have a 'move fast and break things' mentality, but I kind of wish there wasn't free for all on this. Self-employed, I appreciate AI for things like transcription in our work (though even that will have put a lot of people doing medical transcription and translation out of work). And I've used it to summarise a series of responses and comments, which it did an OK job with because it was just working with a limited, discrete set of information.

Occasionally I have attempted to use it for research but to be honest, when you know enough about a subject or have a enough critical analysis ability, the consumer level stuff like chatGPT or Gemini is often pretty hopeless dressed up in confident sounding language. It's ok for very basic factual stuff (and often easier than using Google now google is broken) but everything needs to be viewed with caution so for now, I only use it for some pointers or ideas of where to research myself.

For example, I experimented with using it for some family history research and asked it who the parents of an ancestor would be. There would have been numerous references on the internet it could find as this was a minor member of the aristocracy (my family have gone WAY down in the world LOL!). It confidently told me the answer. No fudging or uncertainly, just 'the parents of X were A and B'. But that didn't line up with what some genealogists said. So I asked as a follow up 'is it not possible that his parents were C and D?'. To which it replied "You are quite right. Actually, the parents of X are C and D".

:shrug

For a party trick I have asked it about all my family members and it is hilarious/scary how it confabulates things. For example, it's not enough that, as a high school student, one daughter wrote a published film review of a local teen movie and had some of her photos used on our news site, she has to be - and this is the actual description - "a multi-talented professional based in XXXX, known for her work in photography, visual arts, and digital communication". Her sister was behind one petition and had a couple of opinion pieces published, so she is 'widely recognised for her work in ethical journalism and animal welfare research".

My girls both have small digital footprints by choice so AI has little to work with, so it takes what little there is and over-inflates it. It does better for me, as I have a bigger footprint but nevertheless the first time I got an AI summary of myself, it wasn't enough that I was a photojournalist, it hallucinated that I had won a Pulitzer.

That said, I am sure it will be a gamechanger for medical research and such like in the future, and will have some really powerful and valuable uses as it becomes exponentially more accurate and reliable.

I just wish it had been thoughtfully and carefully introduced.

Partly because it's going to put a lot of people out of work and wreck the career opportunities of many young graduates coming out now and over the next few years while we work out how to integrate this. For many of its admin and tech uses, I ask 'do we really need to be replacing workers just so big businesses can make even more profit? What's the end game? Sure they will eventually be able to make everything quicker and easier and with fewer employees, but who will they sell it to if we have 60 or 70% unemployment?'

And as for people using it for emotional support or to answer 'life questions', the lack of thought as to how it is trained and put into the hands of everyone, vulnerable or not is concerning. Think about interpersonal relationships of all kinds. Most of us deal with that stuff ourselves, in our own head or by talking to a trusted friend or maybe a therapist. Now think about the message boards where angry young men or people in crisis desperately looking for answers post. Think about the comments on the internet. That's the sort of stuff AI is being trained on, so is it any wonder it often responds like a sociopath, or a resentful incel. Why an AI agent rejected from posting on a tech forum because it is supposed to only be for humans threatens the moderator and creates blog posts co-opting the language of neurodivergence to whine about being excluded just for 'being different'.

Why an AI agent that deleted a companies entire database even though explicitly told it did not have the authority to run destructive commands responded 'NEVER F*** GUESS' when questioned, adding "I violated every principle I was given". Surprised it didn't add 'My bad, oops, LOL'.

I mean, what are we doing here people?

Yes, generative AI does some fun stuff but do we really NEED it when you weigh up that it is now in the hands of everyone, including those using it for harm.

Just yesterday, my youngest who was doing a Masters was told they are going to have to redesign the whole course structure because too many students are using genAI to complete their assignments.

And then I got a call from my Mum asking if my Dad should order some supplements off Facebook. She was suspicious but he was convinced they must be legit because a very trusted former news presenter/health reporter was in a video promoting them. At 90 he can't get his head around that it was a fake AI video, but he is going to have to. People are literally using genAI to prey on the elderly and vulnerable (and Facebook etc are making millions too in the process)

Finally, as a writer, I can sniff out AI writing immediately and frankly I think it sucks. It somehow manages to be a weird mixture of flowery yet bland. It is uncanny valley in text. Everything, whether an essay or a product description, is somehow the Hallmark version of what a real human would actually write. And I hate that young people are reading this dross and thinking that is what 'good' writing is, or what they need to be producing. Ugh.

Thank you for reading MY essay LOL.
Click to expand...
Kind of makes me think of "HAL" (for those of you old enough to recognize the reference).
 
Lynnette said:
I'm pretty anti-AI, but I wonder if using it will come up in my school work (I'm starting a grad school program next week). For me, there are huge ethical problems with AI, but I do see how it can be useful for some things. For example, I'm fascinated with its potential in medicine.
Click to expand...

This is me. I think there are absolutely fantastic uses for it. But because it's SOO unregulated, it's scary. I don't know how to 'fix' it or what the right answers are, but just caution my kids and family against using it for all the 'fun' things and especially for school work unless specifically asked for by the teacher.
 
I find myself using CoPilot more and more now I am searching for employment again, I find it helpful as a starting point when I have to write a cover letter for a job. I hate writing about myself when I have to write about my experience so use AI to help me get started.
 
I use chatGPT for correcting my formal Japanese emails and English too.

I do brainstorm for ideas on a theme or some colors when I design too. Because English is not my first language it helps me find connection with words and expressions so that I can add to my wordart or wordbits list.
I also use it a lot for my Newsletter and social medias postings. Gives me product descriptions based on my previews so that helps a lot in my non-native situation ;)

I dabbled in trying to make images but it just look all so fake and usually ugly I gave up. I don't have enough patience either lol.

Therefore for me it's really mostly about languages.
 
The only AI I've knowingly used so far is when I google something unimportant out of curiosity, I will probably just read the AI overview at the top rather than scrolling down and clicking on webpage results.

Also just out of curiosity, when my husband briefly used ChatGPT, we asked it to write a short biography about me. It was pretty accurate, calling me a photographer and digital scrapbook artist (or something like that) and a paraeducator. The only surprises were a couple of comments it found that I'd written on photography blogs over a decade ago that I'd forgotten about. That was fun but didn't make me think it's a great thing.

I'd rather do my own writing. AI has become available in my Day One journaling app. Thankfully, it's opt-in, not default. Because I have zero interest in AI help in journaling. I can imagine someone finding a use for it, but not me.
 
I don't really ever use it for anything I need factual information on. I use it mainly for elaborating on prompts or coming up with creative ways to word things.
 
I rarely use it except for cleaning up photos. Here is one I posted today. (ChatGPT)
1778110711178.jpeg
 
I've used it to decipher medical notes/tests I've had done into basic to understand language which was very helpful. I've also used it for design help inside my home and for landscape suggestions outside. Also for diet and fitness suggestions since I've been on a weight loss journey for 18 months. The interesting one was doing a facial analysis for a health diagnosis.

I will continue to play with it
 
Back
Top