Email or username:

Password:

Forgot your password?
Artemis's posts Post Back to profile
Top-level
Artemis

Generative AI is making our language and communication poorer. It does this not just by outputting its own garbage but also because now we have AI making arbitrary determinations of what words and phrases we *humans* are or aren't allowed to use.

24 October at 17:44 | Open on dice.camp
31 comments
Artemis

Tell you what: "devoid" was definitely a word I knew and used in elementary school.

Am I normal? Nah, not really.
Am I human? I'm pretty damn sure I am, yeah.

24 October at 17:45 | Open on dice.camp
Artemis

Imagine coming up with a machine that effectively gets to decide how humans are "supposed to" write and speak.

24 October at 17:54 | Open on dice.camp
Hubert Figuière
The Doctor

@artemis Isn't that the public school system?

24 October at 18:30 | Open on hackers.town
Molly Cranberries-Kraig 🦃

@artemis

Reduction to a mean pollutes the source for everyone. The idiocracification of humanity continues apace.

24 October at 18:31 | Open on mastodon.social
Natasha Nox 🇺🇦🇵🇸

@mckra1g @artemis Damn, those are some fancy words. You must be at least 50% AI.

24 October at 22:49 | Open on chaos.social
Oblomov

@Natanox @mckra1g @artemis you can tell it's human because of the neologism. AI aren't able to create new words ATM.

25 October at 2:49 | Open on sociale.network
Natasha Nox 🇺🇦🇵🇸

@oblomov @mckra1g @artemis As a matter of fact LLMs can do that if you instruct them to do so, likewise it can interpret made up words rather well.

25 October at 7:13 | Open on chaos.social
FoolishOwl

@artemis This is the opposite of the curb cut effect. They're harming everyone by repressing autistic people.

24 October at 18:38 | Open on social.coop
Elenna :verified_transgender:​

@foolishowl @artemis "on the suggestion of tech companies we have decided to raise all curbs by 10-20 cm."

25 October at 5:22 | Open on catcatnya.com
Large Heydon Collider

@foolishowl @artemis They're harming everyone directly. It's just automated anti intellectualism.

25 October at 7:18 | Open on front-end.social
A Scape Of Goats

@artemis sounds like a white supremacy machine tbh. prescriptivism has always been racist.

24 October at 19:46 | Open on plush.city
F4GRX Sébastien

@artemis openai will collapse before that.

24 October at 20:25 | Open on chaos.social
RetroEggy

@artemis Feels like some kind of twisted Turing test...

24 October at 20:25 | Open on troet.cafe
Pteryx the Puzzle Secretary

@artemis I've always trended towards a high vocabulary — which, I might add, is something that *used* to be considered an extremely positive thing that was pushed to be grown as much as possible in kids.

And then there's how I basically *had* to go out of my way to learn the word "malicious" back in second grade to counteract some bullshit comebacks my teachers had about the boys teasing me. >_<

24 October at 20:39 | Open on dice.camp
dropbear

@artemis they are addressing the problem in a way that I wouldn't.

Report that to the education institute, up the chain of necessary. I get why teachers want to reduce their workload, it's just the tool not being effective. Scrap it, use a different one, matters not.

Instead of posting on social media for rage bait. I don't even believe they tried to de-AI it, but that is beside the point.

The core of the problem is they try to solve an educational issue with technology, they failed to get their students to be interested in writing and proud of being able to write eloquently; and try to address the cheating with a cat and mouse technology game.

I took the bait, well done.

@artemis they are addressing the problem in a way that I wouldn't.

Report that to the education institute, up the chain of necessary. I get why teachers want to reduce their workload, it's just the tool not being effective. Scrap it, use a different one, matters not.

Instead of posting on social media for rage bait. I don't even believe they tried to de-AI it, but that is beside the point.

25 October at 1:51 | Open on aus.social
Reinder Dijkhuis Does Art

@artemis I got enough of that from my classmates!

25 October at 7:14 | Open on mastodon.art
Tats :disability_flag:

@artemis my mother? let me tell you about my mother

25 October at 7:15 | Open on beige.party
Madeleine Morris
Log 🪵

@artemis I learned what "dearth" meant from Keapon Laffin in Quest for Glory 2. I don't know when or where I learned "devoid" from, but it was definitely before that, and from an equally highbrow literary source.

24 October at 18:49 | Open on mastodon.sdf.org
Brian Tatosky

@artemis Being compared to the average of slop data scraped off the internet in order to be "human" is incredibly awful.

24 October at 17:46 | Open on mastodon.online
Roger Moore

@virtualbri @artemis
It may not even be that. Apparently the AIs need polishing to sound better, which requires a lot of human input. Because Westerners are expensive, they hire people from developing countries to do the work. The AI picks up some of the dialect of wherever the polishing happens, and that's what the "AI detector" notices. If you happen to like using whatever words or phrases the detector has honed in on, it will ding you.

24 October at 23:11 | Open on c.im
Reinder Dijkhuis Does Art

@VATVSLPR @virtualbri @artemis You forgot to mention *highly educated* people from developing countries because that's the population that gets those jobs.

25 October at 7:15 | Open on mastodon.art
PacificNic
Jens Finkhäuser
jnkrtech

@artemis It’s worst in the form of AI but I also hate it in the form of increasingly opinionated “grammar and style” checkers built into every email and doc editing software now. If I couch a phrase in any but the most blunt terms I get little squiggles telling me that I’m communicating wrong.

Computers should not be telling humans how they’re allowed to use language. Humans should not be building programs which crush out variance and expressivity in language.

24 October at 19:14 | Open on social.treehouse.systems
Mikalai

@artemis
Enforcement of stupid rules makes our language and lives poorer.
This time stupid rules are introduced with magic wrap of generative AI.

24 October at 20:51 | Open on discuss.systems
Az

@artemis IIRC there was a huge thing on African Twitter about this: a lot of the LLM training/tagging was done by Nigerian workers (underpaid ghost labour) and Nigerian English is gorgeously full of expressive and specific terms which introduces a bias in LLMs to use those words like delve, devoid, etc, and AI checkers were updated in turn to look for those as markers and people automatically assuming stuff was written by LLMs.

25 October at 2:15 | Open on scorpinc.social
Steffo 🐲

@artemis This is how you could imagine how the robots will take over the humans. They decide what we are allowed to use and what not. And many of us wouldn't notice it. Because it's obviously already happening. :sd_thinking:

Ok, that thought is really creepy o.o

25 October at 6:59 | Open on fellies.social
Clara (Hobo arc) 🍉💕🌈

@artemis Typically I would add a short preamble stating that I am proud of having a rich vocabulary and that I can prove it live.

25 October at 7:15 | Open on eldritch.cafe
Go Up