AI Dungeon vs Have I Been Trained? Comparison

An in-depth comparison of AI Dungeon and Have I Been Trained?

A

AI Dungeon

A text-based adventure-story game you direct (and star in) while the AI brings it to life.

freemiumOther
H

Have I Been Trained?

Check if your image has been used to train popular AI art models.

freemiumOther

AI Dungeon vs. Have I Been Trained?

In the rapidly evolving landscape of artificial intelligence, tools generally fall into two categories: those that help you create with AI and those that help you manage your relationship with it. This article compares AI Dungeon, a pioneer in generative storytelling, and Have I Been Trained?, a vital utility for digital rights and data privacy. While they serve different purposes, both are essential for anyone navigating the modern AI ecosystem.

Quick Comparison Table

Feature AI Dungeon Have I Been Trained?
Primary Function AI-powered text adventure and roleplay. Dataset search and AI training opt-out.
Core Technology Large Language Models (GPT-4, Mixtral, Llama). Image search indexing (LAION-5B dataset).
User Interaction Active: You direct the story and actions. Passive/Protective: You search for your data.
Pricing Freemium (Subscriptions from $9.99/mo). Free (Enterprise API available).
Best For Gamers, writers, and creative hobbyists. Artists, photographers, and privacy advocates.

Tool Overviews

AI Dungeon is a text-based adventure game that uses advanced Large Language Models (LLMs) to generate infinite story possibilities. Unlike traditional RPGs with pre-written scripts, AI Dungeon allows players to type any action or dialogue, which the AI then incorporates into the narrative in real-time. Developed by Latitude, it has evolved from a simple experiment into a robust platform featuring multiple AI models, world-building tools, and multiplayer modes, effectively turning the user into both the protagonist and the director of a living story.

Have I Been Trained? is a public service tool created by the organization Spawning.ai to give creators agency over their digital footprints. It allows users to search the LAION-5B dataset—a massive collection of over five billion images used to train popular AI art models like Stable Diffusion and Midjourney—to see if their work has been included. Beyond simple searching, the tool provides a "Do Not Train" registry, allowing artists to flag their images for removal from future training sets, championing the concept of "consent" in the age of big data.

Detailed Feature Comparison

The primary difference between these tools is their goal: AI Dungeon is built for engagement, while Have I Been Trained? is built for accountability. AI Dungeon provides a "sandbox" experience where you can choose between different AI "engines" (like Griffin or Dragon) depending on how much coherence or creativity you need. It includes features like "Story Cards" to help the AI remember specific lore and an "Adventure" mode where you can share your scenarios with a community of millions. It is a tool designed to consume AI output for entertainment.

In contrast, Have I Been Trained? is a diagnostic and protective tool. Its main feature is a highly efficient search engine that can match uploaded images or text descriptions against billions of entries in AI training sets. Once a user identifies their work, they can use the platform's opt-out mechanism. This doesn't necessarily delete the image from the internet, but it sends a signal to AI developers who respect the Spawning API that the data should be excluded from future model versions. It is a tool designed to monitor AI input for ethical reasons.

From a technical standpoint, AI Dungeon relies on "generative" AI—it creates new content based on patterns it has learned. Have I Been Trained? uses "discriminative" or "retrieval" technology—it finds existing content within a database. While AI Dungeon users are often looking for the AI to "hallucinate" an exciting plot twist, users of Have I Been Trained? are looking for absolute accuracy to ensure their intellectual property is being handled according to their wishes.

Pricing Comparison

AI Dungeon operates on a multi-tiered subscription model. While there is a Free (Wanderer) tier, it offers limited context (the AI's "memory") and uses more basic models. Paid tiers include Adventurer ($9.99/mo), Champion ($14.99/mo), and higher tiers up to Mythic ($49.99/mo). Higher prices unlock "Ultra" models like GPT-4, larger memory banks, and monthly credits for generating AI images within the game.

Have I Been Trained? is essentially Free for individual artists and creators. The search and "Do Not Train" registry are provided as a public service to foster a more ethical AI landscape. However, Spawning.ai offers enterprise-level solutions and APIs (through their Source.Plus platform) for organizations that need to manage large-scale data rights or for AI developers who want to integrate "consent" filters into their training pipelines.

Use Case Recommendations

Use AI Dungeon if:

  • You want to play a role-playing game with no boundaries.
  • You are a writer looking for inspiration or a way to break through writer's block.
  • You want to experiment with different LLMs (like Llama or Mixtral) in a gamified environment.

Use Have I Been Trained? if:

  • You are a professional artist or photographer concerned about your portfolio being used without permission.
  • You want to understand which datasets are powering the current wave of AI art.
  • You want to exercise your right to opt-out of future AI model training.

Verdict

Comparing AI Dungeon and Have I Been Trained? is a study in the two sides of the AI coin. AI Dungeon is the clear winner for entertainment and creativity, offering an unparalleled experience for those who want to interact with AI. Have I Been Trained? is the essential winner for digital ethics and privacy, providing a necessary check on how those very models are built. For most users, AI Dungeon is a toy or a creative partner, while Have I Been Trained? is a shield. Both are highly recommended for their respective (and very different) niches.

Explore More