Radio Hour: You Game of Thrones'd Yourself

It's all wins as Chase and Thomas convene for Episode 39 of the Rascal Radio Hour. Three notable tabletop RPGs were decorated with awards and a hall of fame induction at Origins Trade Show, and United Wizards of the Coast officially elected the first every union within the venerable publisher.

The pair run down the rest of the headlines in a decidedly more optimistic week for the hobby and its attending industry before discussing the games each have been playing recently. Thomas' Band of Blades group are reaching the end of their campaign with a satisyfing, but also suprising, execution. Chase bring part two of his ongoing Diplomacy affair, which sets up either a glorious or disastrous part three.

The Question Dungeon takes the form of a deadly political soiree as both crew members cross s(words) with user-submitted queries. Someone asks for half-year predictions on what novel games the industry might produce; another listener wonders why the concept of dungeon crawling has such staying power in our imagination.

Here's an excerpt:

Thomas: Slightly Reckless Games is not some huge studio. Berserkr is not necessarily a super successful game, but the the studio has an [anti-generative AI] policy. They just crowdfunded a game called Devil Division, which is the first potentially large or rather full-scale hack or reworking of MCDM's Draw Steel. It uses the Draw Steel engine to tell a kind of shonen anime, 1990s' action sort of game. They said they don't use generative AI at any part of the company. So, you know, when these questions come up, specifically what people noticed what was in the text of the game, there is a lot of "it's not just X, but y." If you look up Wikipedia's page on identifying LLM writing, for example, you'll see a lot of different sentence structures, and this is one. And it doesn't appear like once or twice. It's a recurring pattern.

People were like, hey, why is this weird pattern showing up? And some of these sentences are not good as well, right? People are pointing out that saying something like Frigg isn't just a combatant, she's inevitable. What does that mean? Who thought Frigg, the goddess from North Mythology, was a combatant? Who thought runes were just a force and not a language?

We reached out to Slightly Reckless Games as well, and they made a clear statement saying that they did not use AI. They said that every sentence was written by them. The the founder of the company is the main writer. They've said very clearly that they didn't use it. And functionally, for us, that's the end of the line in terms of us reporting this story, right? If they say we haven't used it, fine. We can't really write about it at this point. There is no good way of proving there is no evidence for this allegation. It's possible that it's not.

We're not gonna put out an article that says this allegation was made because even that's problematic in a different way, right? We can't normalize a constant reporting of AI allegations when there is no clear evidence or consensus. We can't run this text through an AI text checker. Even though I've heard from some people that they're reliable, I've heard from a lot of other people that they're not. The question of false negatives and false positives abound. So, what are you going to do? That just puts us in a weird place. This industry, especially the independent side of it, cares a lot about not supporting or using any generative AI in any way.

There are all of these disclosures that have been sort of forced onto platforms. They don't want to even think about this, whether it's Steam or Kickstarter or YouTube. Nobody wants to police this. So, we get these declarations, but even then we don't really know what to do. This is an ongoing problem, and maybe it's the limitations of journalism as an institution about how we think about and report on generative AI in creative industries.

Chase: It's where I see the use of regulatory bodies in other industries. But that does presume that the tabletop industry at large wants to make sure that no AI is used in the the production of games. I think for the small press side, that's true. When you get to the larger publishers, some of them are very adamantly saying "no AI". But others will say there's no AI in their games, but we're not really leading the charge on making sure that this is something that everybody is adopting. It really comes down to individual folks having to say, I don't use AI, and then us believing them.

I remember covering a lot of false positives within Magic: the Gathering and D&D art while at Dicebreaker. There was a lot of like witch hunting online. And I'm not not to say that like we shouldn't try to call folks out, but how do you do it in a way that is responsible and doesn't push small creators, who are already overworked and burdened, to then personally combat any sort of allegations. History has shown that the internet does not exercise good judgment in these matters.

But you're right that there's just so many cases of language that has been unfortunately co-opted by AI as a marker. Listen, we are users of the double dash and the em-dash here. And it sucks that that has become something that tips a lot of people off to AI.

Thomas: Yeah, you'll see in that subreddit thread that people are like, and these dashes, and these dashes, that's evidence. And it's a bit like, come on. There's dashes there because we love using them.

Chase: Yeah, it's because we can't fucking shut up and just use a period before starting a new sentence. We gotta keep going. One more parenthetical.

Thomas: No, exactly. You gotta keep saying things in between other things.