Looking back, I was absolutely oblivious to the scope of the changes that were underway as computers took over, rendering typewriters and library card catalogues extinct.
I still remember my sister Nora swearing that she could not properly edit a manuscript with a computer. Pushing the words, line by line, page by page through the typewriter were essential to a proper rewrite.
Is there ANYONE out there today who actually starts over when they rewrite, "typing" the whole thing over?
My parents wrote a movie (The Desk Set, 1957) which features a gigantic (think: semi truck) mainframe computer matching wits with Katharine Hepburn (and losing). The first PC was 20 years away and today, a computer that powerful can fit into a well ventilated coat closet.
I was sure I'd never swap out my typewriter for a personal computer, my PC for a MAC. Or succumb to the siren song of a cell phone. Or ditch my land line. Or do my banking on line.
Now we're going kicking and screaming into an age of generative artificial intelligence. Griping about "customer service" bots that run us around in circles. Yearning for the voice of another actual human being when we dial HELP.
Meanwhile some folks are singing its praises.
I've been keeping my distance, happily ignorant when it comes to AI tools. Unsure even of what vocabulary to use to describe them. Wondering, like everyone else, if AI will render writers obsolete. Swearing at phone bots and unhelpful "help" bots that seem to make it impossible to tell truth (whatever that is) from a bot's fever dream.
But my pessimism got dented recently.
I had a medical test and waited and waited for the results. After a few weeks of silence, PING, an email came that directed me to click and log into the medical "portal" (of hell?) and see what the tests think.
So, after waiting another week and getting up my courage, I clicked. Searched down the right password (every g-d doctor/practice seems to have their own "portal"). And read results that felt like instructions to a chemist.
There were numbers and percentages and... mixed in, a lot of words. I didn't know whether to be worried. Or relieved. Confused: absolutely. Baffled, for sure.
Then, like a lifeline, the portal offered to run my test results through generative AI.
What's to lose? I clicked the link.
The bulleted list that came up was subtitled "Interpretation." The words were plain and simple, easy to parse. Brief and to the point. One word leapt out at me: Negative. Let's just say that in this case that result was a positive for me.
Maybe I can get used to this new world after all.
I will not be using AI to write or rewrite, and a cautionary tale in last week's New York Times illustrates how a chatbot can feed delusional thinking.
No more than I'll be be hauling out my typewriter and running manuscript edits through it or trying to get Verizon to re-install my land line. But I am a bit less pessimistic about what the future holds and how I'll be adapting to it.
Because, clear as can be, this is the way of the future.
Have you had any positive or negative experiences with generative AI? Or are you sitting there scratching your head and saying, "Generative what??"
Have you had any positive or negative experiences with generative AI? Or are you sitting there scratching your head and saying, "Generative what??"
Feeling very wary of AI in the first place, I've not sought out generative AI, so no pro or con experiences here in that regard . . . .
ReplyDeleteI'm more or less in the bunker with you, Joan.
DeleteI can't help thinking of it as degenerative AI.
DeleteNah, I have been resisting the use of generative AI as much as I can, although I can see how it benefited you in the interpretation of your medical tests.
ReplyDeleteI disabled Gemini on my Samsung phone, turned off Microsoft Copilot and FB's Meta AI, and have never used ChatGPT.
Smart.
DeleteAs for medical test results, Ontario Health is still stuck in the 20th century.
DeleteI got my mammogram, fecal colon test results by letter sent by regular mail this summer!
And I have to phone my family doctor's office to get an appointment. Nothing is online for me in Ottawa.
Grace, I also disabled Gemini, but it keeps nudging its way past my block. Grrr!
DeleteI had a computer professional disable Gemini. What a relief! Any reviews written since mid-July have been written without the nagging misdirection of that most annoying pest.
DeleteYeah, I don't need Gemini or Microsoft Copilot's (in Word) help to write book reviews!
DeleteWell I'm in the camp of scratching my head. I don't think I even know how to contact it or recognize it. I don't think I'm using it but who knows?
ReplyDeleteOops, that was me, Elizabeth.
ReplyDeleteIf you're using Google, then I think you're "using it"... it seems to come up as "AI Overview" at the start of search results.
DeleteAutocorrect is another form of AI.
DeleteKaren in Ohio, thank you. I just asked Annette about that.
DeleteGood, I have shut off autocorrect on my PC, android phone and tablet. That's why there's plenty of typos/spelling mistakes. I can't see the small screen type that well, but I prefer this.
DeleteGrace, I have also turned off autocorrect. :) Selden
DeleteSame for me, Grace and Selden. And since I have a tremor that often affects my typing (especially on the phone) AND can't see without my glasses, some embarrassing mistakes get made. What a pain. But... at least I am not getting "corrected" to misspellings by artificial "intelligence".
DeleteI distrust AI and when I need specific information I have learned to turn off AI in Google searches by ending my query with "-ai". When you really know something about a subject, you realize how basic or even harebrained some of the responses are. I tend to have more trust in my research skills.
ReplyDeleteAbout ChatGPT, I once wrote a scholarly history article. It was extremely dense and layered, with a a great many biographical facts and quotations that hadn't been used before. (It had taken me months to figure out how to scaffold it.) It was published. Another editor asked me if I would use the same facts and quotations in a pair of articles on the two main people covered. I couldn't imagine how to do so without quoting myself. He said cheerfully, "Oh, just feed everything into ChatGPT and see with what it comes up with!" It may have been good advice, but I declined.
Grace, I'd love to know how to disable Meta AI. (Selden)
Grace, if you're out there, a lot of us would love to knw how to disable Meta AI, too.
DeleteHow to turn off Meta AI Chat! (I did this in the FB app on my Android devices, not on my PC).
DeleteFB usually rolls out features so if you don’t see it check again in a few days
1. Hit the arrow next to the chat
2. Hit the i in the upper right corner
3. Hit MUTE
4. Click “until I change it”
Grace, is it possible to turn off the Meta AI Chat on Instagram too?
DeleteDIANA: I don't use iG as much as you do, so I had to look it up.
DeleteHow to disable Meta AI in Instagram app
Step 1. Mute or remove the AI assistant in messages
Open the Instagram app and go to Direct Messages (DMs).
Find a conversation with the Meta AI assistant — it’s usually pinned near the top.
Tap into the chat, then tap the profile icon at the top of the screen.
Select “Mute” or scroll down to choose “Remove chat”.
This won’t fully erase AI from your app, but it will stop the assistant from appearing in your inbox.
Step 2. Turn off AI in search and suggested content
Tap the search bar at the bottom or top of the screen.
If a blue circle appears, long press it.
You’ll get a menu where you can select “Disable AI suggestions”.
Then go to Settings > Search history > Clear everything to reset your suggestions.
I’m sure I am using it without even realizing it. I appreciate knowing how to turn it off or get around it like you have mentioned with the google searches Selden.
ReplyDeleteQuit using Google for sesrches.
DeleteUse Duck Duck Go. This search engine respects your privacy. You need to ask for ai help, or it is not included in the search.
They also do not keep data on you to sell like Google does.
I second. I switched to DDG several months ago and appreciate not being fed the AI slop.
DeleteI hate when AI intrudes. When I try to write an email from my website, the stupid AI button covers the entire text book, so I can't see what I'm writing! No, I do not want a robot creating that email for me!
ReplyDeleteI can see using generative AI to interpret medical tests, though. Still, for a lot of things, I'm not sure I trust its answers to questions. Basically, I'm definitely not using it to write, and for everything else, I'm in a wait and see position.
Yes, why does it seem as if suddenly our laptop screens have shrunk relative to the content we're trying to see. And if AI is so smart, why doesn't it realize that?
DeleteAnnette, is it similar to Autocorrect on your phone? When I type my stories, I turn OFF the WiFi since I think AI uses WiFi?
DeleteDiana, not quite the same, and it's a new update. There may be a way to turn it off, but I haven't taken the time to find it yet. Besides, whatever I do is usually un-done during the next update.
DeleteLooks like I'm in good company here. Googling with -ai appended is helpful. I'll never use Chat GPT to write. I find the offer to let AI take over intrusive everywhere. My doctor of many years adds his own interpretive note in the portal (the Gateway in their case) about test results that is clear and human written.
ReplyDeleteI also registered all my books published before August 2022 with the suit against the pirates that fed them into AI without my permission. Grrr.
Also - glad your results were negative! Such a relief.
DeleteThanks, Edith - me, too. Exhaling for now.
DeleteIt’s been so shocking to see the number of writers who have found that their copyrights haven’t been registered, even those working with Big 5 publishers. Glad you are registered, but such a shame this IP theft occurred.
DeleteEdith, why are you still using google? Switch to Duck Duck Go.
DeleteInteresting Hallie, I'm glad you tried it on the medical test--what a relief! Weeks sounds too long to wait for something like that. Those portals are maddening, though I see why companies prefer them. I have used AI to draft things like press releases. then I fix and rewrite in my my own words. I certainly won't use it to write a book though, what would be the point?
ReplyDeleteI've used it for the same kind of make-work work. Then reworked the result
DeleteI receive the results for things like blood test in a few hours via my chart but have to wait weeks to speak to the doctor about them. If I'm curious about a certain test result (it's all about the numbers, isn't it?) I do a quick search on a reputable website.
ReplyDeleteI'd done that but still didn't trust myself to interpret the result.
DeleteI am equal parts amazed, horrified and terrified...I've used ChatGPT once: I recently had an op ed on the importance of the arts published in our daily newspaper; the French radio station called to invite me to do a live on-air interview about it the next day. Yikes! I can speak French but it's rusty and I no longer owned my 2-volume set of English/French dictionaries. Then it hit me: Run the piece through ChatGPT and see what happens. Well. What happened was that it spit out my article in French in about a nano second. AMAZING. I used that version to brush up on the relevant vocabulary I needed for the interview. And haven't touched ChatGPT since. It is FRIGHTENING to think how invasive and pervasive this tech already is in our daily lives...and maybe especially in the writing world where writers' work is 'scraped' to teach AI what it needs to perform its function.
ReplyDeleteWow - now that use IS genuinely useful.
DeleteMy mild dyslexia and now beginning memory issues have blocked me from writing for a very long time. When Bard -now Gemini - was introduced with Google, I began to use it. Like Hallie, we found it very useful while interpreting Amy's cancer notes. These days I find that I leave out words and don't see it before I hit publish. Y'all might thing it is no coffee consumed morning for me, uh nope. So running it through Gemini to check for spelling and clarity helps my anxiety index go lower.
ReplyDeleteAs an experiment I ran the above paragraph through Gemini asking for a proofread check, Here is the result:
My mild dyslexia and recent memory issues have made writing difficult for a long time. When Google introduced Bard—now Gemini—I started using it. Like Hallie, I found it incredibly useful for interpreting Amy's cancer notes. These days, I often leave out words and don't notice it before I publish. You might think it's because I haven't had my coffee yet, but that's not the case. Running my writing through Gemini to check for spelling and clarity really helps lower my anxiety.
Coralee, I concede that that sounds useful and reasonable. It certainly makes the checking and rechecking simpler and much less stressful. If I could guarantee it only would help in that case, but stop nagging me with words it chooses, having no idea about what I am striving to say, then I would gladly use it for that purpose. But, it seems to be a case of all or nuttin',
DeleteOMG that revision is pretty amazing. I can see how useful that would be if you're writing NONfiction... as long as you check that meanings didn't get subverted. For fiction, though, I wonder if it can handle what we call "voice" -- the idiosyncracies of a particular narrator, imparted to the prose.
DeleteHi ~ Paula B here. Coralee, I prefer your “uh nope”. The rewrite was very professional sounding but devoid of your special voice. No personality. And, blah.
DeleteAn interesting example, thanks, Coralee! And I agree with Paula B--as I was reading it I wanted more of "you" than the AI generated. Which is as it should be--you are unique, and no machine can recreate that.
DeleteI agree with Paula. This is what I want in informal writing. Some people might be surprised that I often guess who is commenting here by their words and the way they use them. Yes, I know the names are there but since I scroll from the bottom I don't see the names until I am ready to.
DeleteI love this game! Thank you for telling us about it.
DeleteIn terms of the test results, your story leaves me feeling very fortunate. I go in today for my annual physical, and did my blood work last week. First I feel fortunate because all of my results posted to the portal WITH A REFERENCE range of what was considered normal; second, because my own doctor posted his quick summary of his opinion about them only a day later. But then, I assume your tests, Hallie, were ordered by a specialist, and they are renowned for their lack of sensitivity to patients' feelings and concerns. I'm so glad the results, when you finally got them, were good!
ReplyDeleteAs for AI, count me among the avoiders. I will admit that I don't mind the "AI Overview" in Google when, as is often the case, I'm looking for a specific point of fact -- something like "What year was this song released" or "What are common over-the-counter medicines for this?" But I am grateful to now know how to turn it off when I'm looking for something more indepth. And I immediately followed Grace's clear instructions to turn on Meta AI, which is nothing but annoying. (Thank you Grace!) Don't even get me started on the frustrations of AI "Help" bots that run one around in circles!
I agree on the unhelpful AI "Help" bots. They seem designed to KEEP you running in circles. I'm waiting for companies to start seeing it as a PLUS if they have real-people HELPers. Just look at how popular Apple stores are.
DeleteHALLIE: As Susan said, in terms of your test results, that is good news for you. My Uncle was one of the first people to use a computer. Back in the 1950s, the computers looked like the computers in the movie Desk Set and he showed my grandparents the computer lab at the University. I remember in the 1980s, he showed me how to use a computer with a modem. However when I tried it at home, it was not a success. That is the difference between a computer genius and someone who was not familiar with computers.
ReplyDeleteAt University, I remember going across the street from campus to the typewriter store and renting a typewriter to type my papers for classes because our professors required that all papers be typed! Renting a typewriter was within my student budget while computers were way too expensive. Even Apple computers at that time were too expensive. It was not until after I graduated that Apple started selling user friendly computers with the MacBooks. I am typing this on my Apple laptop.
Regarding AI, the other day I saw the headlines about a hospital in ? South Korea ? Germany ? using AI to diagnosis brain tumors! I have mixed feelings about AI. I use voice to text on my phone to check on my speech. Some of the letters are more challenging for me to pronounce, although my cochlear implant speech processors help me hear the slight differences. I think it is "muscle memory" for me. Practice makes things better.
However, I REFUSE to use AI nor ChatPT?GR? to help me write my stories. I already have many ideas for stories. On another note, someone (a Jungle Red) is launching her novel tomorrow on Tuesday the 9th of September. I think I will see a post tomorrow about the book.
Diana, you're reminding me that I wrote my PhD dissertation on an electric typewriter. With copysets. On the other hand, I used a mainframe to analyze the data. A foot in both words.
DeleteHallie, I remember using carbon copies because our professors told us to make sure we had two copies of every paper we submitted to class.
DeleteI really like the idea of using AI to evaluate test results. I have turned to WebMD instead, which really increased my anxiety.
ReplyDeleteI have mulishly refused to read articles about AI or use it. However, the AI attached to Google's search engine really helped me recently when I looked for arguments against a congressional resolution (the innocuous sounding Defend the Dams act--how bout calling it Kill the Salmon, but I digress). AI came up with a lovely bullet-point list of arguments. I was impressed. Perhaps it has some helpful applications!
Oh, Gillian, those misnamed bills pushing everything from vaccine avoidance to strip-mining our National parks and forests (okay, it's a little hyperbole, but not much) make me crazy. Good job, you!!
DeleteAI is big on bullet points
DeleteI don't really know much about AI, except I use it daily to ask questions of Google answered by AI. I think AI will greatly improve science and medicine outcomes, actually AI already is.
ReplyDeleteThis sounds like something AI generated... kidding.
DeleteHaha. No I'm a real person. I did disclaim that I'm not very knowledgeable about the subject though!!
DeleteHallie, thank heavens for your good results! I too have a medical portal where I can see my results, with a range of normal values posted with each test. My doctor is good at calling to let me know if they need to see me or whether I can go another year for my next tests.
ReplyDeleteI have used Chat GPT twice--once to update my resume--it did a great job of giving me one I could tweak. I've mentioned on here that I've self-published several novels. I don't have a publishing house behind me to create great cover art, and I don't have the budget to hire an artist. If I'm lucky I can find a pre-made cover that will suit the story (and that I can afford), but otherwise it's up to me to create a cover. For one novel--a historical suspense--I wasn't very happy with the cover I'd done. That book consistently undersold compared to another historical suspense novel I'd published. My nephew suggested I try ChatGPT and I did. It took me a while to prompt and prompt and prompt, but I finally ended ended up with a cover that works. The new cover has definitely made a difference in sales.
All I can say is that I really like to know my sources when I research anything. If the Google summary is summarizing most(?)all(?) sources, how do I know they are reputable? I guess I’m a skeptic first, ally second type of person.
ReplyDeleteAmen.
DeleteHallie, sigh of relief for negative results! Whew! Good.
ReplyDeleteI know the tension of waiting for very specific medical results and how much I want to have them immediately or sooner. Fortunately, I have mammogram results sent directly to my GP and my GYN. Each office calls me to report the results within a day of them posting to my medical portal, which I must sign into. The company that does the testing, follows up with a letter.
The doctor who performs the colonoscopy comes into the recovery room and details what he has or hasn't found. It is very efficient. If there needs to be a biopsy, and so far there has not, I expect that he will phone with results before the letter arrives with test results.
My GP follows up blood tests with discussions of whether I should start to take additional meds.
As for other pushy AI bots, I avoid them. I had a pro disable a couple of them on my computer and although they are not totally gone, they are enough in the background to be less of an issue. What bothered me most is not that it wanted to spell things for me ( although we all know that "autocorrect" is happy to change a word here and there...gr-r-r), but that it wanted to anticipate what I was trying to say. The unmitigated gall!
Oh, and don't get me started on telephone portals that spin you in circles and then hang up on you. Who designs those? Who requests that sort of design? Would a firing squad be too extreme?
Oh, Judy, I approve of the firing squad!
DeleteDebRo
The telephone portals are the worst!!!!!!
DeleteAggghh.
Just thinking about it making me upset!
Ever since i got some wrong information about my family tree at Ancestry, I have been very leery. I knew the information they "suggested" couldn't possibly be correct and then I saw the disclaimer that warned that AI was used for some findings. But how many people just willy-nilly accept that information as absolute truth and then present it to others as accurate? Guess it is best to check more than one source, but even then I have doubts about things I find.
ReplyDeleteI suppose, if I were writing a story, it could be useful to have AI invent an entire family tree for a character. I'm thinking that this is something the villain might do, but I've no idea why, other than covering his tracks. Sooner or later someone will probably come up with that idea if they haven't yet.
Judi, I use Ancestry all the time. I find that more than 50% of the suggestions made are incorrect. However, I'm most typically researching 18th c. people who are not famous. (Selden)
DeleteJudi, that is a great idea about the villain using AI to invent a family tree as a cover up for a more serious criminal action? I use Ancestry myself and since I used the DAR Library in person years before the pandemic, I know some of the information is fact. I do not use AI for genealogy research.
DeleteProbably most of what Ancestry tells us is fact, but not all. At least they warn us that some findings were the result of AI. They said that my great grandmother died in Tampa FL when I know for a fact she never left the county. So perhaps someone else with a similar name and birthdate was the cause for confusion.
DeleteIs nothing sacred? (With AI, apparently not...)
ReplyDeleteI was about to sign up to listen to r Audible's versions of Middleworch, then I got a message from the ai "assistant,: asking if I had questions, eg, what's the plot of Middlemarch? Put-leeze--it' snot the plot it's the writing.(Plus, this will be my third time listening to the novel so I know the plot!) Oh, dear. George Eliot would be turning over in her grave.
Hallie, it does look like I'm using it. Certainly in the ways you mentioned. Still scratching my head, though.
ReplyDeleteNo one was happier than me when the pc was invented. I hated typewriters and that couldn’t type as fast as I could think, made mistakes and had to use whiteout.
ReplyDeleteIn the past weeks AI has been driving me bonkers. With a new book out I’ve been getting non stop letters to help me promote it. They all begin with an AI assessment of the story. Readers will appreciate the brave decisions made etc!
one of the great things about using the pc for writing is that the words automatically move to the next line and there is no need for whiteouts. However, I turn off the WiFi when writing so there is no chance of AI stealing my information online.
DeleteAs a tool, well, thank goodness for the help it offers for things like interpreting test results (that we used to be able to rely on via our doctors), or for Coralee's dyslexia, or for the sight-impaired, etc, etc.
ReplyDeleteHowever, it's a Pandora's Box of ills, too. Now that the Box has been opened, there is no containing the ills it's released into the world. And that worries me.
Also Hallie, much gratitude for good test results for you. Hugs.
DeleteAgreed! Due to their iterative nature, bad results are now part of the training data for the next round, and so on, and so on . . .
DeleteGood on the negative result, Hallie. Glad that there is at least one good use for AI. Not a fan otherwise, except for spell check. As for AI customer service help bots....Dante, we need a new ring for them.
ReplyDeleteAgreed! They are their own special circle (which is what I feel like I am going in when forced to interact with them).
DeleteI don’t know if Siri would count as a Bot but I’ve found Siri to be less and less helpful over the past few years. A couple of months ago I asked a question and Siri’s response was “I’m sorry, but I don’t know anything about that.” I don’t remember if I ever found the information I was looking for.
ReplyDeleteI’m sure that there are instances in which I have unknowingly used AI or Bots, or whatever, but I don’t really like their uninvited presence in my life.
DebRo
Siri's response is hysterical!! A least it gives a good laugh.
DeleteWhat was your question?
I don’t remember the question, and I can’t remember if I ever found the information!
DeleteI don't use Siri or Alexa or whoever. I happily ignore any extraneous remarks online when I'm reading emails or looking up things. As for when I'm writing, if a word pops up in anticipation I can use it or lose it.
ReplyDeleteLast month at WorldCon, I got to participate in a table talk with Emily M Bender, one of the co-authors of THE AI CON (which I just finished reading). The TL;DR is that generative AI can be helpful when it has been developed for a specific application. That usually means it has been trained on a narrower dataset that has been (hopefully) well-vetted.
ReplyDeleteThe scarier ones are the gen AI that are trying to be everything to everyone. These models are trained on so much data scraped from all over the internet. Think about some of the dumbest things you’ve read - yep, that’s going into the model. So many of these models are just language predictors - by percentages, what is the next most likely word to go in this sentence? Bender refers to them as “stochastic parrots.”
For searches, I much prefer to see the context of the results. I can see the top 10 responses and pick and choose which I will read and trust rather than just accepting what is thrust at me.
I also have huge issues with the social and environmental issues associated with training and using these models. Spending my career in water and power utilities, trying to find every way we could increase conservation - it hurts to see it all being thrown away. So many communities are going to find their systems, which may have already been behind the curve on reinvestment, faltering under the increased water and power demand from data centers. And the human capital to train the models and remove erroneous or hateful content is mostly in the global south and occurs in very questionable working conditions.
[Stepping off of my soapbox]
Hallie, I am very happy for your positive results. Hopefully gen AI can be developed to increase this kind of utility while reducing the negative aspects.
I think AI has its uses...like showing us cats doing Olympic sports! :-) but I get annoyed when AI offers to write for me. No thank you, I'll do my writing all by myself!!
ReplyDeleteI'm not a writer but thought I'd respond to this. I've not tried AI due to it's uploading your chats and other information. I just read that DuckDuckGo has a deal with ChatGPT where ChatGPT can't upload chats and other personal information. I'm going to try it out soon. It's either available at duckduck.ai or by going to duckduckgo and searching for it.
ReplyDelete