hagenone argument i hear in favor of ai generated text is »i’m not a writer, this will make it sound better/professional« and i get it. there is a reason i’m not sharing sketches i draw on the internet. they look like shit. but honestly: i’d much, much, much rather read something »imperfect« but that sounds like a human with an actual personality and emotions wrote it than the wall of beige that falls out of chat gpt. my shitty drawings are more interesting than midjourney images. embrace humanity.
»but sometimes my mails are too emotional for a workplace setting« fuck this! be emotional. be mad at somebody. apologize after. have honest interactions with people. be cringe. be everything a chatbot isn’t. i can’t deal with this void of nothing that falls out of chatgpt.
FJ!!@hagen Sorry, but this is elitist. I know people having to sell their house because they haven't been able to be hired in a year.
"Ask forgiveness later" is something
a) white cis-het men get to do, everyone else gets warnings and HR visits
b) once they are already hired, because you don't get to apologize for an imperfect cover letter and CV once it is in the wastebasket.
@fj i would agree that applying for a job is an exception to what i said.
Reemt@hagen if I get a support request by the sister companies CEO in corporate English, I use ChatGPT to draft my response. I’m confident in my English abilities to a certain degree. But corporate lingo and the specifics of replying to a French native speaking English in Hong Kong… I’d rather take all the help available to me.
@hagen The modern workflow for translators since 2017 or so has been to take the original text, put it through DeepL, and then meticulously check the result for mistakes. This is actually much faster than having to do the 90% drudgery of translating yourself, and thus allows translators to do more of their already underpaid work so they can make a living.
Note: I never said let DeepL work unsupervised. Good translators do not do that.
Are you telling these people to either go slow again, and watch their livelyhoods drop, or to that you'd much much rather read an imperfect translation of business documents because communication needs to be human?
TeflonTrout 
he/him@hagen Wall of what? Slander
bedast@hagen What also concerns me is some neurodivergent folks tend to be very mechanical with their written/typed words and it looks and sounds like it’s from ChatGPT. 20 years ago I was writing emails and getting positive remarks about how I write them. Today, those same emails would be thought as made with a LLM. The false positives are showing more and more, never mind the detection tools just being bad.
njsg@hagen I kinda think that expecting GenAI to sound professional only works if the bar for that is set a bit low. Might be realistic in some fields or places, though - I mean, GenAI "professional" is the text version of pixelated graphics with compression artifacts being used for logos in official documents...
@njsg i don’t mean professional as in like a professional writer. every job needs some writing.
gkrnours@hagen professional writer don't work alone, they have editor who help them make their world look professional.
@gkrnours thank you for explaining to me, a professional writer, how professional writers write.
@hagen sorry, that wasn't the intention :( I wanted to point that the fact of using genAI to do alone a work that's usually not done alone isn't the best idea and one shouldn't be expected to have writing that look professional while working alone