My ear worm this week was Aretha's "Rock Steady" and while listening to it on repeat I wanted to learn about the making of the song. Google AI brought up information about "Respect."
Michael Jensen is a quirky, sixty-year-old travel writer and digital nomad who has been exploring the world for the past seven years. He's known for his off-the-beaten-path adventures and his humorous yet reflective commentary on travel and life. Michael's daily routines often include morning walks to immerse himself in the local culture, which he then documents through his engaging writing and photography.
His personality shines through his work, characterized by a mix of thoughtful introspection and a knack for uncovering the unusual. Michael is also deeply interested in the connections between history and place, bringing fresh perspectives to well-trodden destinations while championing hidden gems. He has a talent for weaving personal anecdotes into broader narratives, making his stories relatable and resonant.
Michael's recent projects include a Substack newsletter that pairs his striking travel photography with his unique take on different destinations. His past destinations include a wide variety of locations, from Sarajevo and Athens to Bangkok, showcasing his adventurous spirit and willingness to embrace the unknown.
This is funny, Ramona. I’ve been doing these kinds of tests of chatbots for a couple of years, because I’m concerned about the ethical quandaries for writers this all brings up. In my case, it could draw generic info about me because of Talking Writing that was mostly accurate, even right after OpenAI first released a public version of ChatGPT at the end of 2022. It could also mimic my writing style, although the bots now have more guardrails in place to (supposedly) avoid such ripoffs. (Ha!)
When I tried your experiment today, it came back with stuff from my website, titles of my books, and a few other sources: Martha writes about first-person journalism, mental-health issues, adoption, etc. Boring - and you’re right, it is a suck-up, designed to be a helpful customer service agent giving you what you ask for.
When I asked it to list some of my specific articles, it generated a whole raft of them from Salon in the wayback days - which only indicates how search algorithms, SEO, and what’s freely available play a part in the information generated. It’s one of the most problematic things about doing research with bots - we can fact-check obvious hallucinations but we also need to question what’s missing. When it comes to print work that hasn’t been digitized (which applies to many books and smaller journals pre-1990), it can feel like whole archives have evaporated.
What it tells me is that I'm invisible to algorithms, and that's not funny. I don't know what to do about that, seriously. What I need is a PR person! I'm terrible at it. I simply cannot blow my own horn or convince anyone to read me, let alone hire me.
I agree, being invisible to algorithms is serious — and it's going to become more serious as corporate media and tech companies control those algorithms. But I don't think what a chatbot dredges up necessarily reflects the total impact or visibility of a writer, especially one who is as active as you are on Substack. I have an increasing sense that "writers are doing it for themselves" these days (including journalists), and part of doing it is to recognize each other and to read each other's work.
ChatGPT doesn't recognize me at all, saying I'm not widely recognized in mainstream literary circles or databases.I'm shocked. Shocked, I tell you! Perplexity, on the other hand (https://www.perplexity.ai/) mined content directly from Substack and various work-related bios, including a few photos - two that weren't me.
I should have also added that I appreciated reading what GPT brought up for you. Seemed to be more interpretive than Perplexity (see below, minus the bulleted formatting that didn't transfer). It's an open-source AI model that I found via Jane Friedman. If I remember correctly, she liked it better than GPT.
Ramona Grigg is a versatile professional writer with extensive experience across multiple writing domains. She is a:
Freelance Feature Writer: Professionally writing feature articles for newspapers and magazines14
Columnist: With a long-standing career writing columns45
Political Blogger: Ran a political blog called "Ramona's Voices" for 10 years, starting on Barack Obama's Inauguration Day57
Creative Nonfiction Writer: Focuses primarily on creative nonfiction writing through her Substack publication "Writer Everlasting"2
Essayist: Writes essays on various topics including life, love, loss, politics, and humor3
With over 40 years of writing experience, Grigg has also been a Creative Writing teacher and writers conference speaker12. She currently maintains two newsletters on Substack: "Writer Everlasting" and "Constant Commoner", where she shares her writing and explores diverse topics2.
AI CANNOT THINK! Maybe if I repeat it enough, the Algorithm will hear me. AI can't think. It can't analyze. It can't interpret. It doesn't know ANYTHING. All it can do is string the most predictable sentences together, using its search engine, to appear as though it can answer a question. Dr. Holly Walters on BlueSky 5-25-24
No, of course it can’t. It can only cut and paste, which is why we writers are so against it. It grabs from us and uses our words without shame or compensation. 😏
This was simply a fun test to see what it had on me.
I am absolutely STUNNED by this and obviously a bit flattered but, um, I am definitely not Australian lol! I wonder if this is because my generous friend over at knotsintheString, Prue, has mentioned me a few times?
Sue Sutherland-Wood is an Australian writer known for her literary fiction and creative nonfiction. Her writing often explores themes of personal identity, family dynamics, and the complexity of human relationships. She tends to write with a focus on introspection and the emotional undercurrents that shape her characters’ lives.
Her work might be categorized as contemporary or reflective fiction, often diving into the inner lives of her characters while exploring broader societal themes. If you're interested in emotional depth and character-driven narratives, her work could be a great fit.
I tried asking Claude (yet another chatbot) about "the writer Jack Herlocker" and got Sorry, who? as a response. "I recommend seeking verification through specific sources like library catalogs, book review sites, or direct contact with the author or publisher." <sigh>
Ramona, I wouldn't put too much weight on these results. ChatGPT, and other so-called generative AI, are only as smart as whatever they're trained to learn. If no one of its handlers reads Ramona Grigg to it, it doesn't know anything about her writing. AI is also prone to inventing "facts" out of whole cloth, possibly misguiding people about your work, even if it was introduced to it.
That AI is so prone to error and outright fictional invention, and people are still so invested in it being the future, is absolutely beyond my comprehension.
You can correct Chat AI, but when you ask again later, it will spiel out the same made up answers. One time Chat AI told me I was famous for inventing the term used to describe birds that get mulched in jet engines. No such term existed, it was all made up. The AI developers must be laughing their asses off over there.
Yeah, ChatGpT will not criticize or tell you the truth. You have to know the right questions to get good answers. Like I told it to give me three sentences and and avoid cliches and it gave me cliches’. When I corrected it, it gave me even worse cliches. Lol
So this was a lot of fun to read, and makes me want to do it for myself and see what nonsense they come up with! I’m a dummy—how do you do it? What link?
They got a lot of it right, but got overly attached to the idea that everything I do is somehow about gender and the body!! Also got some titles totally fucked up:https://chatgpt.com/.
So this time it came up differently, with many more mistakes about what I’d written and what’s in the books. I screenshot for awhile, but then got knocked off the site:
My ear worm this week was Aretha's "Rock Steady" and while listening to it on repeat I wanted to learn about the making of the song. Google AI brought up information about "Respect."
I think I'm going to have to learn some restraint here. This could become a new addiction! Proving it wrong might be way too much fun.
Here is what I got:
Tell me about the travel writer Michael Jensen.
ChatGPT said:
ChatGPT
Michael Jensen is a quirky, sixty-year-old travel writer and digital nomad who has been exploring the world for the past seven years. He's known for his off-the-beaten-path adventures and his humorous yet reflective commentary on travel and life. Michael's daily routines often include morning walks to immerse himself in the local culture, which he then documents through his engaging writing and photography.
His personality shines through his work, characterized by a mix of thoughtful introspection and a knack for uncovering the unusual. Michael is also deeply interested in the connections between history and place, bringing fresh perspectives to well-trodden destinations while championing hidden gems. He has a talent for weaving personal anecdotes into broader narratives, making his stories relatable and resonant.
Michael's recent projects include a Substack newsletter that pairs his striking travel photography with his unique take on different destinations. His past destinations include a wide variety of locations, from Sarajevo and Athens to Bangkok, showcasing his adventurous spirit and willingness to embrace the unknown.
TL:DR: I'm QUIRKY!!!
I love it! It’s so YOU! But they found your Substack and couldn’t find mine?? Either of them?? I’m mortified… 🙄😆
It's ChatGPT! I would neither be mortified nor care because it's not all that! LOL.
This is funny, Ramona. I’ve been doing these kinds of tests of chatbots for a couple of years, because I’m concerned about the ethical quandaries for writers this all brings up. In my case, it could draw generic info about me because of Talking Writing that was mostly accurate, even right after OpenAI first released a public version of ChatGPT at the end of 2022. It could also mimic my writing style, although the bots now have more guardrails in place to (supposedly) avoid such ripoffs. (Ha!)
When I tried your experiment today, it came back with stuff from my website, titles of my books, and a few other sources: Martha writes about first-person journalism, mental-health issues, adoption, etc. Boring - and you’re right, it is a suck-up, designed to be a helpful customer service agent giving you what you ask for.
When I asked it to list some of my specific articles, it generated a whole raft of them from Salon in the wayback days - which only indicates how search algorithms, SEO, and what’s freely available play a part in the information generated. It’s one of the most problematic things about doing research with bots - we can fact-check obvious hallucinations but we also need to question what’s missing. When it comes to print work that hasn’t been digitized (which applies to many books and smaller journals pre-1990), it can feel like whole archives have evaporated.
What it tells me is that I'm invisible to algorithms, and that's not funny. I don't know what to do about that, seriously. What I need is a PR person! I'm terrible at it. I simply cannot blow my own horn or convince anyone to read me, let alone hire me.
I agree, being invisible to algorithms is serious — and it's going to become more serious as corporate media and tech companies control those algorithms. But I don't think what a chatbot dredges up necessarily reflects the total impact or visibility of a writer, especially one who is as active as you are on Substack. I have an increasing sense that "writers are doing it for themselves" these days (including journalists), and part of doing it is to recognize each other and to read each other's work.
ChatGPT doesn't recognize me at all, saying I'm not widely recognized in mainstream literary circles or databases.I'm shocked. Shocked, I tell you! Perplexity, on the other hand (https://www.perplexity.ai/) mined content directly from Substack and various work-related bios, including a few photos - two that weren't me.
Better keep the day job.
Wow, that's pretty awful. How did you find out? And what is Perplexity, anyway? I went to their site but couldn't figure it out.
I should have also added that I appreciated reading what GPT brought up for you. Seemed to be more interpretive than Perplexity (see below, minus the bulleted formatting that didn't transfer). It's an open-source AI model that I found via Jane Friedman. If I remember correctly, she liked it better than GPT.
Here's the Wiki page: https://en.wikipedia.org/wiki/Perplexity_AI
_______
Ramona Grigg is a versatile professional writer with extensive experience across multiple writing domains. She is a:
Freelance Feature Writer: Professionally writing feature articles for newspapers and magazines14
Columnist: With a long-standing career writing columns45
Political Blogger: Ran a political blog called "Ramona's Voices" for 10 years, starting on Barack Obama's Inauguration Day57
Creative Nonfiction Writer: Focuses primarily on creative nonfiction writing through her Substack publication "Writer Everlasting"2
Essayist: Writes essays on various topics including life, love, loss, politics, and humor3
With over 40 years of writing experience, Grigg has also been a Creative Writing teacher and writers conference speaker12. She currently maintains two newsletters on Substack: "Writer Everlasting" and "Constant Commoner", where she shares her writing and explores diverse topics2.
Wow! I feel seen! Thanks. (It looks like they found my Linkedin page.)
AI CANNOT THINK! Maybe if I repeat it enough, the Algorithm will hear me. AI can't think. It can't analyze. It can't interpret. It doesn't know ANYTHING. All it can do is string the most predictable sentences together, using its search engine, to appear as though it can answer a question. Dr. Holly Walters on BlueSky 5-25-24
No, of course it can’t. It can only cut and paste, which is why we writers are so against it. It grabs from us and uses our words without shame or compensation. 😏
This was simply a fun test to see what it had on me.
I am absolutely STUNNED by this and obviously a bit flattered but, um, I am definitely not Australian lol! I wonder if this is because my generous friend over at knotsintheString, Prue, has mentioned me a few times?
Sue Sutherland-Wood is an Australian writer known for her literary fiction and creative nonfiction. Her writing often explores themes of personal identity, family dynamics, and the complexity of human relationships. She tends to write with a focus on introspection and the emotional undercurrents that shape her characters’ lives.
Her work might be categorized as contemporary or reflective fiction, often diving into the inner lives of her characters while exploring broader societal themes. If you're interested in emotional depth and character-driven narratives, her work could be a great fit.
Nice. Be sure and correct it.
I tried asking Claude (yet another chatbot) about "the writer Jack Herlocker" and got Sorry, who? as a response. "I recommend seeking verification through specific sources like library catalogs, book review sites, or direct contact with the author or publisher." <sigh>
Well, that bot doesn’t know what it’s missing! 😍
I’m sorry, Ramona, but based on that bland, dull, cliche summary I would NOT read your work. It sounds frightfully boring, lol.
Yeah, that’s probably why I’m not a household name… 😏
Ramona, I wouldn't put too much weight on these results. ChatGPT, and other so-called generative AI, are only as smart as whatever they're trained to learn. If no one of its handlers reads Ramona Grigg to it, it doesn't know anything about her writing. AI is also prone to inventing "facts" out of whole cloth, possibly misguiding people about your work, even if it was introduced to it.
That AI is so prone to error and outright fictional invention, and people are still so invested in it being the future, is absolutely beyond my comprehension.
I guess I should have used sarcasm emojis. I wrote this for fun only. As a writer I have no use for AI and I’ve said so often.
Oh gosh, my time's too precious. AI is a threat to everything creatives hold dear, IMO.
I think we can all agree on that!
You can correct Chat AI, but when you ask again later, it will spiel out the same made up answers. One time Chat AI told me I was famous for inventing the term used to describe birds that get mulched in jet engines. No such term existed, it was all made up. The AI developers must be laughing their asses off over there.
Yeah, ChatGpT will not criticize or tell you the truth. You have to know the right questions to get good answers. Like I told it to give me three sentences and and avoid cliches and it gave me cliches’. When I corrected it, it gave me even worse cliches. Lol
Ha! It’s good that it makes stupid mistakes. If it didn’t we might be tempted to use it. 😉
So this was a lot of fun to read, and makes me want to do it for myself and see what nonsense they come up with! I’m a dummy—how do you do it? What link?
Just go to https://chatgpt.com/. It should take you right to the place where you can ask a question. Have fun! I'd love to know what comes up!
They got a lot of it right, but got overly attached to the idea that everything I do is somehow about gender and the body!! Also got some titles totally fucked up:https://chatgpt.com/.
Your link just goes to the main page. Can you cut and paste?
OK.
I can’t seem to get back to anything except the first paragraph.
Maybe ask again?
So this time it came up differently, with many more mistakes about what I’d written and what’s in the books. I screenshot for awhile, but then got knocked off the site:
Don’t see how to post screenshots here in comments.
Not screenshots, just cut and pasted text.
Interesting that it came up differently. Clear evidence that it's worse than unreliable.
What a fun post, Ramona. Thanks!
It's a Large Language Model, not a search engine.