Embracing AI in TTRPGs - Enhancing, Not Replacing, Creativity

[fluffybunbunkittens]

The fact that these two statements are made, one right after the other, has made this post absolutly worth it.

Because individual people is not the same as corporations..? Creating a sample image of your character for friend group use is not the same as putting them on the cover of a book you're selling?

 

[Theory of Games]

However, a number of studies show that people are using AI in large numbers, they just wont admit it

Link to those "studies"?

 

[OptionalRule]

Link to those "studies"?

We've officially departed the "let me google that for you" world and entered the "Let me ask AI that for you" world.

I have asked AI for you. The answer is below. I look forward to your artful use of ironic "quotes" in future replies.

___

While there have not been any definitive large-scale studies showing widespread covert AI use, some research provides insights into this topic:

## Reported AI Usage in Education

A survey by ACT found that 46% of high school students in grades 10-12 reported using AI tools like ChatGPT for school and non-school assignments[1]. However, this only accounts for self-reported usage, not unreported use.

Similarly, a survey by BestColleges found that 56% of college students admitted to using AI on assignments or exams[2]. Again, this relies on students' willingness to disclose their AI use.

## Potential Underreporting

There are indications that actual AI usage may be higher than reported:

- In the ACT study, students with higher test scores were more likely to report using AI tools, suggesting some students may be reluctant to admit usage[1].

- The BestColleges survey showed a significant increase in reported AI use compared to a previous survey, jumping from 22% to 56% in less than a year[2]. This rapid rise could indicate growing acceptance in admitting AI use.

## AI Use in College Admissions

A study by foundry10 found that about 33% of high school seniors applying to college in 2023-24 acknowledged using AI tools for help with admissions essays[3]. However, only 6% admitted to using AI to write entire essay drafts.

## Challenges in Measuring Covert AI Use

Accurately measuring undisclosed AI use presents challenges:

• Reliance on self-reporting may lead to underestimation due to social desirability bias.
• AI detection tools are not foolproof and may miss sophisticated AI-generated content.
• The rapidly evolving nature of AI technology makes it difficult for studies to keep pace.

While these studies provide valuable insights, they likely do not capture the full extent of AI use, especially covert usage. More comprehensive research methods may be needed to accurately assess the prevalence of undisclosed AI use across various contexts.

Citations:
[1] Home
[2] 56% of College Students Have Used AI on Assignments or Exams | BestColleges
[3] 1 in 3 College Applicants Used AI for Essay Help. Did They Cheat?
[4] Balancing the potentials and pitfalls of AI in college admissions
[5] 8 in 10 Colleges Will Use AI in Admissions by 2024 - Intelligent
[6] [7] [8] https://www.usnews.com/education/best-colleges/articles/is-ai-affecting-college-admissions

 

[Scribe]

We've officially departed the "let me google that for you" world and entered the "Let me ask AI that for you" world.

Yes, when glue is suggested as an addition to Pizza?

I'm reminded of when my son was able to Google his Math homework. His scores improved, did he actually learn anything?

 

[Umbran]

For corporations, sure. But ethics left the room well before AI was developed when it comes to corporations.

There is still a personal component.

Your game books do not have enough content to train an LLM sufficiently to give cogent English responses - so you're taking a package already trained, and pointing it at a local source for focused information. And exactly what data was that prior training on?

It is a thorny question - how far removed do you have to be from an unethical source for your use to be ethical?

The worst takes are clearly conrolling the discussion space, at least online.

I mean, I can understand that you feel they are the worst takes. But this is not a decided issue - history will tell what the worst take was. You and I don't really get to do that here and now.

However, a number of studies show that people are using AI in large numbers, they just wont admit it. It's worth a discussion, even if people are afraid to talk about it.

"Everyone is doing it!" is not an ethically powerful defense of an action.

Maybe people should be a little afraid (or maybe embarrassed is a better term) to talk about it, much as they should be embarrassed by having a public conversation where they crow about how snazzy and expensive their sneakers are, while ignoring the working conditions of those who sewed them together.

 

[Umbran]

We've officially departed the "let me google that for you" world and entered the "Let me ask AI that for you" world.

Hey, if you want to make an assertion, you have to back it up. This passive-aggressive "let me google that for you" is trying to put the burden of proving your point on someone else.

You don't get to shame people over not proving your point for you, even jokingly. That's uncool.

 

[OptionalRule]

Hey, if you want to make an assertion, you have to back it up. This passive-aggressive "let me google that for you" is trying to put the burden of proving your point on someone else.

You don't get to shame people over not proving your point for you, even jokingly. That's uncool.

No. You've come in telling people they should be afraid. Making blanket statements about ethics and that everyoen has to share that or it's uncool. I get it. You're a mod. You can do what you like. But it doesn't make you right.

 

[kigmatzomat]

As a person with minimal artistic ability, if there was an art AI that was trained exclusively on out-of-copyright material I would use it for personal projects. But it doesn't exist because curation requires effort and no one wants to do that work when they can scrape the whole internet and pay lawyers instead.

By the same token, I am baffled no one has released GutenbergLLM. There should be enough stuff in project gutenberg to get a copyright-free English language LLM. Add a dozen pages of modern words and uses to give it some extra context. Bonus points if there is public domain government documents that can be fed to it to provide some additional modern context.

And I am of the opinion transformer output should not be copyrightable or patentable.

 

[OptionalRule]

As a person with minimal artistic ability, if there was an art AI that was trained exclusively on out-of-copyright material I would use it for personal projects. But it doesn't exist because curation requires effort and no one wants to do that work when they can scrape the whole internet and pay lawyers instead.

By the same token, I am baffled no one has released GutenbergLLM. There should be enough stuff in project gutenberg to get a copyright-free English language LLM. Add a dozen pages of modern words and uses to give it some extra context. Bonus points if there is public domain government documents that can be fed to it to provide some additional modern context.

And I am of the opinion transformer output should not be copyrightable or patentable.

There are a number of different LLMs built around these concepts right now. The majors aren't one of them of course, but there's a good discussion happening around this that's worth participating in.

Also, you can't copyright anything produced with LLMs. Courts have ruled on this already.

 

[Umbran]

No. You've come in telling people they should be afraid. Making blanket statements about ethics and that everyoen has to share that or it's uncool.

I have raised questions about the ethics. I've noted where there are thorny issues here. I have also admitted that I don't get to decide what the wrong takes are, any more than you do.

Do you contend there are no ethical questions around generative AI?

I get it. You're a mod. You can do what you like.

Ha! I wish.
No, I don't actually wish. But no, I cannot generally do what I want.

If nothing else, at the most basic, I answer to Morrus, who has a vested interest in this place operating well. I have no doubt he'd show me the door if I got in the habit of doing whatever I want, unless what I wanted was pretty darned nice.

But, more important (at least to me) is that I feel there is a Spider-Man Principle in action in this role.

But it doesn't make you right.

I already noted that these are OPEN questions - which admits that I don't know I am correct.