This is one of those moments where paying attention changes what you do next. AI roleplaying games and interactive fiction deserve more careful attention than the typical coverage provides, and the reason is not complicated once you know where to look.
Here’s what most people miss: Latitude’s AI Dungeon hit 2 million users within months of launching on GPT-2. When you look at the numbers, this tells us everything we need to know about where things are headed.
The Criticism: Setting the Terms
AI Dungeon launched in 2019 and became the first mainstream AI-driven RPG with millions of users. That’s not just a data point, it’s the foundation that makes everything else in this space make sense. These conditions have been building for years, and what we’re seeing now is different from previous moments that might have looked similar.
Here’s what happened: AI Dungeon hit those 2 million users fast, then KoboldAI and NovelAI emerged as alternatives with local model support and fewer content restrictions. Look at both together and you’ll see a pattern that AI Dungeon has been dealing with from the inside. The conditions are more solid than they first appeared, and the implications go way beyond the immediate headlines.
To understand why this matters, compare what was true three years ago to what’s true now. The change isn’t just about numbers, it’s about quality. The players, the infrastructure, the incentives have all shifted in ways that build on each other rather than cancel out. That compounding effect is what you need to track.
What makes this worth examining carefully isn’t the novelty but the confirmation. The underlying dynamics have been visible for a while. What’s new is that they’ve reached a point where ignoring them takes real effort rather than simple inattention. Crossing that threshold is the real event here.
And Character.AI reaching 20 million daily active users largely through roleplay? That’s part of the same picture. These aren’t separate developments, they’re reinforcing conditions in the same shift.
The Honest Review: The Analysis
Character.AI’s 20 million daily active users mostly doing roleplay is where things get specific. The surface reading isn’t wrong, but it misses the mechanism. And the mechanism is where the useful insight lives.
What people miss: tabletop RPG communities are using LLMs as dungeon masters for solo play sessions. Understanding that changes everything about how you read this information.
This isn’t some random correlation. It’s a direct result of structural factors that have been building up. Previous attempts to understand similar situations failed because they treated symptoms as causes. The structural view is less exciting as a headline but more useful for actually understanding what’s happening.
Comparing this to prior cycles is helpful precisely because of where the comparison breaks down. Similar-looking conditions played out differently before because the foundation was different. AI-generated branching narratives are challenging traditional visual novel and CYOA formats, and that represents a foundation change. It alters how flexible the system is, not just its current state. Recognizing that distinction separates real analysis from pattern-matching.
The skeptical view deserves honest engagement: previous moments with similar surface characteristics didn’t produce the outcomes that seemed logical at the time. That history is real. What’s different now is that AI-generated branching narratives are challenging traditional visual novel and CYOA formats. This isn’t a minor variable, it’s the infrastructure condition that previous cycles lacked. Infrastructure changes tend to stick around in ways that sentiment-driven changes don’t. NovelAI creative writing is tracking this dimension with the attention it deserves.
There’s also a distribution question that often gets ignored in coverage of AI roleplaying games: who captures the value from these shifts, and who absorbs the disruption costs? The big picture can be positive while the distribution is uneven in ways that matter enormously to specific participants. Keeping that lens in view is part of reading the situation clearly rather than just optimistically.
Implications: What This Means If You Care About New releases
The implications extend beyond the immediate context. AI Dungeon’s launch and mainstream success, combined with the structural conditions I’ve described, creates a situation where adjacent fields, decisions, and communities get affected in ways that aren’t always visible from inside the primary story. The second-order effects are often more important than the first-order ones.
Here’s where this perspective differs from mainstream coverage: KoboldAI and NovelAI emerging as alternatives with local model support and fewer content restrictions is a leading indicator, not a lagging one. The people positioned to respond to what this signals, rather than to what it confirms, will be less surprised by what comes next.
Your practical response depends heavily on where you sit relative to these dynamics. For those closest to the core of AI roleplaying games, the implications are immediate and operational. For those at greater distance, the implications are strategic. It’s about understanding which adjacent pressures are building and which assumed stabilities are more fragile than they appear.
The practical question isn’t whether to engage with these dynamics but how. The answer depends on context, on what role you occupy and what your actual decision timeline looks like. But the first step is the same regardless: accurate understanding of what’s actually happening rather than what the most available narrative says is happening.
A few concrete observations worth separating from the broader analysis. First: AI Dungeon hitting 2 million users within months on GPT-2 is not a temporary condition, it’s a new baseline. Second: tabletop RPG communities using LLMs as dungeon masters for solo play suggests the adjustment period isn’t over. Third, and most important: the organizations and individuals treating the current moment as a new steady state rather than a transition are making a classification error that will be costly to unwind later.
The Case Against: What the Critics Get Right
Intellectual honesty requires acknowledging the strongest counterarguments, not just the weakest ones. The case against the optimistic reading of AI roleplaying games isn’t trivial. There are structural vulnerabilities in the current picture that deserve direct engagement rather than dismissal.
The most serious objection is about sustainability. KoboldAI and NovelAI emerging as alternatives can be read not as a foundation but as a ceiling. A point beyond which growth becomes self-limiting because of the very dynamics that produced it. If the current state has already absorbed most of the available early-adopting participants, the remaining growth curve may be structurally shallower than the recent trajectory implies.
There’s also the policy and regulatory dimension. AI Dungeon’s success describes a condition in a relatively permissive environment. Regulatory responses to the scale implied by these numbers aren’t inevitable, but they’re not implausible either. Organizations planning as though the current regulatory environment is permanent are making an assumption that the history of fast-growing sectors doesn’t support.
The rebuttal to these concerns isn’t that they’re wrong, it’s that they’re already partially priced into the current state of the field. AI-generated branching narratives challenging traditional visual novel and CYOA formats reflects an environment where participants are already adapting to constraints rather than operating in an unconstrained space. The adjustment capacity of the ecosystem is higher than a purely top-down view of the risks suggests.
Looking Forward
The trajectory here is clearer than the pace. Making predictions about when specific thresholds will be crossed is genuinely difficult, and anyone claiming precision about timelines should be treated with skepticism. But the direction toward continued development of the conditions I’ve described is supported by evidence in a way that doesn’t depend on a single variable going right.
AI-generated branching narratives challenging traditional visual novel and CYOA formats is the variable to watch as the leading indicator. Historical patterns suggest it moves first, with broader metrics following with some lag. This doesn’t make the outcome certain, but it makes it readable. And readability is the precondition for good decisions.
Three questions are worth holding as the story develops. First: are the structural conditions that enabled the current state durable, or are they cyclical? Second: who is positioned to benefit from the next phase, and does that differ materially from who benefited in the current phase? Third: what would a clean falsification of the optimistic thesis look like, and is there any evidence of that signal emerging? These questions don’t need answers today, but having asked them changes what you notice in the months ahead.
The action from here is straightforward, even when the situation isn’t. The current moment in AI roleplaying games is one where people who have built an accurate model of the underlying dynamics are better positioned than those relying on the surface story. Building that model isn’t quick, but it’s doable. This analysis is intended as one input into it.
If you’ve put time in, how does this track with your experience?