[…] said the company has now been able to mitigate the serious mental health risks and […]
Wow. So they don’t just ignore all the warnings about detected suicidal thoughts any longer? As far as I remember from the news stories, that specific teenager got flagged hundreds of times by their systems for spiralling towards suicide. It’s just that OpenAI didn’t do anything…
That’s going to be a disaster. It’s going to overly reinforce toxic sex understanding due to lack of sexual education. Most porn already does that, but is limited by what a human body can actually perform. AI will be able to make what is essentially sexual torture look like it’s what someone should normally expect sex to be like and make it look like the recipient is enjoying it…among many other issues.
People have been using AI for this for a while now. I pretty much stopped contributing to AI Horde because almost everyone who used it was using it to generate taboo (incest/bestiality/torture/kiddie stuff) sex roleplays.
Yeah, but this will make it much easier and using much better algorithms to make it more realistic because OpenAI will start profiting off of it and devoting resources to it which they have way more of than the existing systems.
and its concepts for what is acceptable or not will definitely be some bigoted bullshit
I mean, Ao3 has existed for years, and literotica for years longer, but okay, if someone wants something personalised. Vague concerns of plagiarism aside, it’s interesting. The problem I see is that people will take it back to those sites and upload it as their own and not declare it as AI.
We already have AI generated music that has fooled some people. I think AI generated stories will be harder to track. And it’s going to be divisive. One thing I’d like to see is someone prolific like Stephen King do it as an experiment. So King — or, more likely, a young intern working for him — prompts an AI to write a book in the style of Stephen King. And they prompt it to include certain things. Then, Stephen King himself actually writes the book, and the two are packaged together for the cost of the one book. You get the Stephen King version, and you get the AI version which has some ridiculous name. And it’s going to tell the same story, but it’s going to do it slightly differently. So you can read King’s version — I think most readers would either read this one first, or this one only — and then, if you’re up to it, you can read the AI version. This would do two things. One, it would (possibly) prove that AI cannot fully replace human writers. Two (and not mutually exclusive to the previous point), it would give you an alternate-reality version of the first story, and that could be interesting.
This would do two things. One, it would (possibly) prove that AI cannot fully replace human writers. Two (and not mutually exclusive to the previous point), it would give you an alternate-reality version of the first story, and that could be interesting.
this is just “imagine if chatbots were actually useful” fan-fiction
who the hell would want to actually read both the actual King story and the LLM slop version?
at best you’d have LLM fanboys ask their chatbot to summarize the differences between the two, and stroke their neckbeards and say “hmm, isn’t that interesting”
4 emdashes in that paragraph, btw. did you write those yourself?
A recent winner of the Akutagawa prize in Japan said she used chatbots to write around 5% of her novel so they’re already proving useful.
Are you asking how to make em dashes? Hold dash on a phone, Shift+Option+Dash on a Mac, Alt+0151 on Windows.
Shit, you might be able to blame me for LLMs using em dashes. Been using them forever. Maybe even longer than you’ve been alive.
Or, and hear me out, King could collaborate with another human writer to do the exact same fucking thing to promote human creativity instead of giving AI any more publicity.
Well no shit. You could also maybe stop a tank if you stand in front of it, but maybe it’ll just keep going.
I mention King because he did a similar thing before. He wrote two books, one as himself and one as his alter ego Richard Bachman, and the two books — Desperation and The Regulators — mirrored one another. He also was one of the first nationally/internationally published authors to support ebook (he released a book online only).
A recent winner of the Akutagawa prize in Japan said she used chatbots to write around 5% of her novel.
After 33-year-old writer Rie Kudan won the Akutagawa Prize last week, she told reporters that a small portion of her book, Tokyo-to Dojo-to (Tokyo Sympathy Tower), was lifted verbatim from ChatGPT.
“This is a novel written by making full use of a generative A.I.,” Kudan said in her acceptance speech, according to the Japan Times’ Thu-Huong Ha. “Probably about 5 percent of the whole text is written directly from the generative A.I. I would like to work well with them to express my creativity.” https://www.smithsonianmag.com/smart-news/this-award-winning-japanese-novel-was-written-partly-by-chatgpt-180983641/
What about publishing a collection of short stories, some of which have human authors and others from LLMs. You could call it, “2 truths and an AI”.
It’s impressive how they have changed the expectations around their pro it as time goes by:
-Our AI could become sentient and destroy humanity to become sole ruler of the world
-Our AI will discover the cure for cancer and so many other illnesses
-Our AIl will soon have PHd level intelligence
…
-You can sext with our AI
This is the feature that ChatGPT lacked that made me switch to Grok for a while which didn’t seem to have any of these restrictions. I switched back though as ChatGPT just seems to perform much better overall. And it’s cheaper.
