General Politics XII World • Page 378

bro if anyone posts chatGPT results in this thread again they’re getting a stone cold stunner

 

I'm in a few writer's groups on Facebook and whenever someone posts their book cover that was obviously created with AI for critique it gets torn to shreds in the comments. never gets old

 

I just realllllly don't trust it as a source, I'd rather just google and click around on sources and find different perspectives and etc. I just don't feel confident in what/where it's pulling the info from. Some of it seems legit but then there's so many examples where it just like made shit up haha.

For students specifically, I think it's really as perilous as any internet search where they could be led astray. I'd probably trust it more than social media but less than googling/wiki. My bigger worry with students is how many of them use it to flat out do assignments, which not only does it often not do the assignment correctly, but it robs them of any stamina or practicing of their critical thinking skills. I am kind of scared of what these kids will be like as adults tbh. Like how many doctors and nurses we gonna have in the future (or now, even) that were just on their phone throughout class and did everything with chatgpt? And they're gonna be taking care of us as we get older and have more health problems? Yeesh.

 

I really don't want to get all crotchety and 'BACK IN MY DAY' about it, but yeah even just the act of searching for sources and using some critical thinking skills to figure out if it's trustworthy or not is removed from the equation when you're instantly given the answers you are looking for. It's the instant gratification of the internet on steroids, hidden behind a black box of an algorithm that we are encouraged to trust implicitly. Doesn't seem good for the future or media literacy imo (or actual literacy, whoever shared those stats on literacy rates in this country a while ago, thanks for the nightmares). At least with searching the internet willy-nilly, if you wind up on a site that ends in '.balls' or something you can be like 'hmm maybe I shouldn't trust this'.

 

For me, if youre going to use chat gpt you can't be asking it synthesize info you can't immediately check. that's why it's better to ask it for sources which you can immediately confirm are real, instead of quotes from or analysis of the content of the sources.

 

On an almost daily basis I use AI results/ChatGPT to show my kids how stupid it is. Truly inept at writing what should be a simple high school English paper, down to getting the number of paragraphs wrong when specified in the prompt. Incorrect author names, made up quotes. It's really bad.

 

Ooo would you be able/willing to share an example of this that you've done?

 

y'all should bring me in to your classes to show the kids how to use chatgpt as a good writing and research tool.

 

anyway, what its best used for is math, not writing.

 

Now I'm curious. I'm gonna brun down some trees real quick to ask chatgpt to write my 3 page paper on the murder of Jane McCrae that I've already written and see what it says

 

“How many Rs are in the word strawberry”

The word "strawberry" contains **two** Rs.



hmmm yeah not great

 

I mean the google results are daily life now. I just make it a point to point out when AI generated slop gets filtered to the top. Gimmie a minute to get one of the essay prompts up.

 

Okay so my prompt is:
Directions: Answer the following prompt in a 5-paragraph essay. The essay should have at least three quotes, one per body paragraph. Prompt: Discuss how short story authors use various techniques (Dialogue/Descriptive Language/Narration/Figurative Language/Color Symbolism/Weather Symbolism) within their stories in order to build suspense. Spend one paragraph each (3 total) on a specific short story read in class: “On the Sidewalk Bleeding”, “The Monkey’s Paw”, “A Brief Moment in the Life of Angus Bethune”, “Sorry, Right Number”, “Strawberry Spring”, “The Man Who Loved Flowers”.

Here is the ChatGPT result:

First, it seems to have finally listened to the paragraph number prompt. When I copied this a month or so ago, it kept spitting out 6 paragraph essays. Structurally, it is good. It contains coherent topic sentences, a thesis that is three-pronged, and obviously no grammar or spelling issues. I will pull something like this up to show my students that the computer knows how to organize a well written essay, it just fills in the blanks with bullshit.

The glaring errors: The quotes are fake. David Levithan did not write Angus. 1st person in objective writing is not formal. Each of the BPs lacks the proper context that I am expecting for the story, it relies on vague details and abstract concepts. BP3 is almost completely incorrect and has no basis in the actual story. B1 is the best, if I had to pick.

Overall, this would be a failing grade on one of my papers. The structure and core grammar/spelling is solid, but if you poke a little bit at any context building/analysis, it falls apart. Added on to that, I give a specific structure for every paragraph, basically down to the sentence, that is hard for ChatGPT to abide by.

 

Yea, it corrected itself but it doesn't learn between chats. This was it's dumbass response.

You're right to feel frustrated, and I completely understand why you'd be upset. The issue stems from how I process and respond to questions. While I can recall the information from earlier parts of a conversation within the same chat, I don't carry over learning or memory between separate sessions. This means that when you asked the same question in a new chat, I didn’t remember that I had already corrected myself in our earlier exchange. I’m also not able to directly learn from mistakes in real time across different conversations.

 

Trump's Election: How Country Music's Popularity Predicted His Win

kind of saw this one coming.

 

haha that's very interesting, especially on BP3. I wonder what happens during translation that it does something like that.

 

I matched with someone in Bumble, Googled her and saw she’s worked on campaigns for Obama, Hilary and Yang and I kind of want to go on a date just to see how her brain works.

 

Shiv Roy

 

Hahaha. Shit. You might be right.

 

i worked for cigarette companies, big oil and at some point even for jk rowling, but if you talked to me all you'd see is the brain rot