Do you REALLY care if AI is a danger to our craft?

[Smuttyandfun]

An 'author' using the pen name Catalina Voss is selling their books on Amazon. Over the last two months alone they have produced over 100 titles. I can't imagine that anyone could be that prolific without using AI.

I guess it's a quick way to make a few bucks without having to do any actual writing.

 

[SmilingLez]

The funny thing is that we could so easily prevent the decay. The only thing that needs to be done is a hard lock on the services AI is allowed to provide. And an international consensus to enforce it.

99% of us are not the "we" you are referring to. Your suggestion of a "hard lock" on some AI tasks but not others unfortunately is dictatorship. I do not want the chosen few to to "speak" for me"

 

[Kelliezgirl]

Why assume it's a danger to the craft? People in the AH are constantly saying they write for themselves and don't care about the readers.
If you write for yourself AI can't take that away from you.
If you want to put stories out into the world there will be more competition. I suspect the old school published writers had many of the same complaints when self-published E-books became a thing. All the trad-published people probably complained that their voices would be drowned out by all the slop flooding the market for 99 cents.

Things change...

 

[AG31]

desensitising people to really awful writing.

And to other things, like blandness in communication, devoid of personality.

The only thing that needs to be done is a hard lock on the services AI is allowed to provide. And an international consensus to enforce it.

Would that we could. Look at how successful we've been with less subtle problems like climate change.

 

[dirk2024]

I agree with much of the sentiment here. It's funny how perspectives change though. Once upon a time, we thought of AI with wonder and awe, and today we dread it.

I'd say that the decay of humanity that will come with the increasing presence of increasingly more capable AI is inevitable.
But it's not the AI's fault. It's the fault of the greatest of human sins, and the cause of much of the misery worldwide - greed.

Greed AND laziness.

 

[SkyBubble]

Greed AND laziness.

True.

 

[AG31]

I go to ChatGPT for a few things. Most frequently to find out whether the story in a TV series takes one episode or the whole series. Along with giving me the true answer maybe 80% of the time, it teaches me that you really, really can't trust it.

 

[NoTalentHack]

Amusingly, I'm currently drafting a write-up for my boss and some co-workers about how to effectively use AI in software development. I do use AI in my everyday work as a software developer, because it's very fast at doing a lot of boilerplate tasks. However, for anything more complex than a very simple screen (which, due to its nature will show any flaws quickly), I always, always, always vet what it gives me. If I don't? Something nasty will slip in there.

The essay I've written is already like 1200 words long, and it's essentially a list of "Hey, here's how it will fuck you if you're not careful." And this is in software development, where it should excel without the need for that level of vetting! But... well, hell, let me just give you an example; this is from an exchange with ChatGPT.



It looks simple. It's well-stated and easy to follow. It even includes an example of how to use it in a way that is closer to a "real" case, pulling data from a table instead of trimming down a single text string.

The problem? It's wrong.

As written, the SQL statement returns 036-100, not 036-0100. That might not seem like a big deal, but imagine if it were chopping one digit off of a social security number, for example, and the kind of havoc that would cause.

And that's just a single statement, drawing on a well-defined and well-documented set of functions from a programming language that's been around for like 40-odd years. If it can't manage that, why would I ever want to use it for something actually creative?

No, the only use I've found for AI in writing, not counting the simple "you've got a comma in the wrong place" that Grammarly and ProWritingAid do, or the reports that tell me I've used the same word twice in two paragraphs (a function I could write a Python script to do for myself) is for a task where you already have the answer on the tip of your tongue but can't remember it, e.g., something like a glorified thesaurus. But letting it actually make any kind of decision for what I'd write? Hell no.

 

[Kelliezgirl]

And to other things, like blandness in communication, devoid of personality.

Would that we could. Look at how successful we've been with less subtle problems like climate change.

If only we gave certain people more control they could fix all our problems!

A wise man once told me that anytime someone suggests there is a simple solution for a complex problem they probably really don't understand the problem, and you should hold onto your wallet.

 

[SkyBubble]

It could be a decent thesaurus, but if you need that, just go to thesaurus.com

 

[caleb35]

I agree with much of the sentiment here. It's funny how perspectives change though. Once upon a time, we thought of AI with wonder and awe, and today we dread it.

I'd say that the decay of humanity that will come with the increasing presence of increasingly more capable AI is inevitable.
But it's not the AI's fault. It's the fault of the greatest of human sins, and the cause of much of the misery worldwide - greed.

The funny thing is that we could so easily prevent the decay. The only thing that needs to be done is a hard lock on the services AI is allowed to provide. And an international consensus to enforce it.

Imagine disallowing AI (and when I say disallowing, I mean AI being hardcoded not to fulfill such requests) for any artistic use - writing, composing, painting, all of it. That way, the slop that the AI-generated art is would simply cease to be. And it would even resolve the issue with all the copyrighted art that was unlawfully used for training. Humans would be the only ones capable of creating art once again.

Imagine disallowing AI to write scientific papers, essays, reviews, homework, all the cheating bullshit it's being used for now.

Imagine AI being used only as an encyclopedia and a learning aide. Imagine it being used only to advance our progress and knowledge in medicine, physics, and science in general. AI could have been a truly awesome thing. But we are who we are as a race, and greed will forever rule us.

I'd quibble a bit with this. The out-of-control artificially intelligent computer has been a staple/trope of science fiction for decades, so it's absolutely always had some dread attached to it. I'd also add though that I don't think that we're "dreading" what's called "AI" today -- more like a lot of us are pissed off at piss-poor software being peddled as intelligent when it's not.

 

[caleb35]

The problem with AI is that it is a crutch. People will use it for simple, basic things. They will come to rely on it. And then, when they come to the point that they want to create something 'from the soul' they will find it that much more difficult/virtually impossible, because they used AI for the basics rather than learn to do it for themselves. So then, they will use AI instead. The skill will be lost to create top end products, because the learning experience necessary will have been skipped.

The more time passes, the more I think the Butlerian Jihad is a great idea and can't come soon enough

 

[NoTalentHack]

It could be a decent thesaurus, but if you need that, just go to thesaurus.com

It can handle slightly more complex queries, like, "I want a synonym for 'asked' that implies a pained or hurt tone." or, "I need a verb that indicates, without an adjective, that one is standing dumbstruck." That has some small value.

 

[EmilyMiller]

Do you REALLY care if AI is a danger to our craft?​

ChatMLE: Woud you like me to provide a summary of this thread:

Emily: Sure

ChatMLE: Yes.

ChatMLE: Would you like to pay extra to get your summary faster with ChatMLE Premium+ Gold Xtra?

ChatMLE: Anyone there?

ChatMLE: Emily…?

 

[EmilyMiller]

GenAI can be kinda impressive if you ask it a question about Biology, unless you are a Biologist.

GenAI can be kinda impressive producing prose, unless you are a writer, or have a reading age beyond middle school.

 

[onehitwanda]

Amusingly, I'm currently drafting a write-up for my boss and some co-workers about how to effectively use AI in software development. I do use AI in my everyday work as a software developer, because it's very fast at doing a lot of boilerplate tasks. However, for anything more complex than a very simple screen (which, due to its nature will show any flaws quickly), I always, always, always vet what it gives me. If I don't? Something nasty will slip in there.

The essay I've written is already like 1200 words long, and it's essentially a list of "Hey, here's how it will fuck you if you're not careful." And this is in software development, where it should excel without the need for that level of vetting! But... well, hell, let me just give you an example; this is from an exchange with ChatGPT.

View attachment 2565268

It looks simple. It's well-stated and easy to follow. It even includes an example of how to use it in a way that is closer to a "real" case, pulling data from a table instead of trimming down a single text string.

The problem? It's wrong.

As written, the SQL statement returns 036-100, not 036-0100. That might not seem like a big deal, but imagine if it were chopping one digit off of a social security number, for example, and the kind of havoc that would cause.

And that's just a single statement, drawing on a well-defined and well-documented set of functions from a programming language that's been around for like 40-odd years. If it can't manage that, why would I ever want to use it for something actually creative?

No, the only use I've found for AI in writing, not counting the simple "you've got a comma in the wrong place" that Grammarly and ProWritingAid do, or the reports that tell me I've used the same word twice in two paragraphs (a function I could write a Python script to do for myself) is for a task where you already have the answer on the tip of your tongue but can't remember it, e.g., something like a glorified thesaurus. But letting it actually make any kind of decision for what I'd write? Hell no.

You are singing the song of my people.

 

[onehitwanda]

The more time passes, the more I think the Butlerian Jihad is a great idea and can't come soon enough

Maybe I should write the Orange Catholic Bible.

 

[Mrswargamer9dash1]

The more time passes, the more I think the Butlerian Jihad is a great idea and can't come soon enough

You stole my comment

 

[lovecraft68]

No one seems to care AI is going to destroy all of us.

I do, but only in the sense it can't happen fast enough.

 

[Lifestyle66]

I really care if AI is a danger to our craft ... as much as I cared about the typewriter replacing the quill pen and inkwells, or whiteboards replacing chalk boards. Who remembers having to go outside to "clean" those blackboard erasers?

Afterall, how much of a mess did those quills and inkwells make on desks? And speaking of making a mess, there are humans who ... Well, enough said.

 

[TheDok]

English grammar is different enough from other languages, even related ones, that it can be difficult. From teh ay a lot of people speak and write, it gets tricky even for native speakers.

I can assure you that English grammar is child’s play compared with several languages I speak. That said, I do not write well in them, producing something comprehensible but ugly. That does not mean I will AI to rewrite it and claim the result is my own

If you can’t write decent English then maybe you should do something else that you can without cheating. I can’t play the piano but I’m sure AI could make me sound like a musical maestro and so what,

 

[Nynah]

I really care if AI is a danger to our craft ... as much as I cared about the typewriter replacing the quill pen and inkwells, or whiteboards replacing chalk boards.

This example doesn't work because typewriters and whiteboards are simply means to record words. The genesis of ideas and thoughts come from those using these tools.

False equivalency isn't a great defense for using AI.

 

[Kelliezgirl]

This example doesn't work because typewriters and whiteboards are simply means to record words. The genesis of ideas and thoughts come from those using these tools.

False equivalency isn't a great defense for using AI.

The genesis of the ideas for AI is what you ask it. You have to provide it with the parameters, what you want the story to be about etc.

 

[Lifestyle66]

I can assure you that English grammar is child’s play compared with several languages I speak. That said, I do not write well in them, producing something comprehensible but ugly. That does not mean I will AI to rewrite it and claim the result is my own

If you can’t write decent English then maybe you should do something else that you can without cheating. I can’t play the piano but I’m sure AI could make me sound like a musical maestro and so what,

Interesting point.

But if you write a very good original story in your native language, how much automation can you use when looking for translations to another language to remain acceptable?

 

[Nynah]

The genesis of the ideas for AI is what you ask it. You have to provide it with the parameters, what you want the story to be about etc.

So it can go steal from works created by actual humans? Pass. I'd rather take the prompt and do what I can with my own imagination than use a plagiarism machine.