AI Dungeon realized the dream many gamers have had since the \u201980s: an evolving storyline that players themselves create and direct. Now, it\u2019s going further with a new feature that enables players to generate images that illustrate those stories.<\/p>\n
Developed by indie game studio Latitude, which was initially a one-person operation, AI Dungeon writes dialogue and scene descriptions using one of several text-generating AI models \u2014 allowing players to respond to events how they choose (within reason). It remains a work in progress, but with the emergence of image-generating systems like Stability AI\u2019s Stable Diffusion, Latitude is investing in new ways to liven up players\u2019 narratives.<\/p>\n
Within the AI Dungeon UI, you can select a \u201cSee\u201d option at any time to prompt Stable Diffusion to generate an illustration.<\/p>\n<\/div>\n
Access requires a subscription to one of Latitude\u2019s premium plans, which starts at $9.99 per month. It\u2019s a credit-based system \u2014 generating an image costs two credits, with credit limits ranging from 480 per month for the cheapest plan to 1,650 for the priciest ($29.99 per month). On the AI Dungeon client available through Valve\u2019s Steam marketplace, which is priced at $30, members get 500 credits with their purchase.<\/p>\n
\u201cWith Stable Diffusion, image generation is fast enough and cheap enough to offer custom image generation to everyone. Image generation is fun on its own, and being able to create custom images to go with your AI Dungeon story was a no brainer,\u201d Latitude senior marketing director Josh Terranova told via email.<\/p>\n
Unlike image-generating systems of a comparable fidelity (e.g., OpenAI\u2019s DALL-E 2), Stable Diffusion is unrestricted in what it can create excepting the versions served through an API, like Stability AI\u2019s. Trained on 12 billion images from the web, it\u2019s been used to generate artwork, architectural concepts and photorealistic portraits \u2014 but also pornography and celebrity deepfakes.<\/p>\n
Latitude hopes to lean into this freedom, allowing users to create \u201cNSFW\u201d images, including nudes, so long as they don\u2019t make them public. AI Dungeon\u2019s built-in story-sharing mechanism is currently disabled for stories containing images \u2014 a step Terranova says is necessary while Latitude \u201cfigure[s] out the right experience and safeguards.\u201d<\/p>\n
![\"AI](\"https:\/\/www.ghanamma.com\/gp\/wp-content\/uploads\/2022\/09\/Screenshot-23-e1662660511541.png\")
<\/p>\nImage Credits:<\/strong> Latitude<\/p>\n<\/div>\n That\u2019s taking a big risk. Latitude landed in hot water several years ago when some users showed that the game could be used to generate text-based simulated child porn. The company implemented a moderation process involving a human moderator reading through stories alongside an automated filter, but the filter frequently flagged false positives, resulting in overzealous banning.<\/p>\n Latitude eventually corrected for the moderation process\u2019 flaws and implemented an acceptable content policy \u2014 but not until after some serious review bombing and negative publicity. Eager to avoid the same fate, Terranova says that Latitude is taking steps to \u201csensibly\u201d curate AI-generated images while affording players creative expression.<\/p>\n \u201cWe are working with Stability AI, the makers of Stable Diffusion, to ensure measures are in place to prevent generating certain types of content \u2014 mainly content depicting the sexual exploitation of children. These measures would apply to both published and unpublished stories,\u201d Terranova said. \u201cThere are several unanswered questions about the use of AI images that all of us will be working through as AI image models become more accessible. As we learn more about how players will use this powerful technology, we expect adjustments could be made to our product and policies.\u201d<\/p>\n In my limited experiments, the new Stable Diffusion-powered feature works \u2014 but not consistently well, at least not yet. The images generated by the system indeed reflect AI Dungeon\u2019s imagined scenarios \u2014 e.g., a picture of a pirate in response to the prompt \u201cYou come across a captain\u201d \u2014 but not in a similar art style, and sometimes with details omitted.<\/p>\n Image Credits:<\/strong> Latitude<\/p>\n<\/div>\n For example, Stable Diffusion was confused by one scene particularly rich in detail: \u201cYou hide in the bushes. You spot a group of thugs, who are carrying a bundle of money. You leap out and stab one of the thugs, causing him to drop the bundle.\u201d In response, AI Dungeon generated an image of a swordswoman in a forest against a backdrop of a city \u2014 so far so good \u2014 but without the \u201cbundle of money\u201d in sight.<\/p>\n Another complex scene involving skirmishing goblins gave Stable Diffusion trouble. The system seemed to focus on particular keywords at the expense of context, generating an image of warriors with bows instead of goblins pierced by a sword and arrows.<\/p>\n AI Dungeon lets you toggle the prompt to fine-tune the results, but it didn\u2019t make a massive difference in my experience. Edits had to be incredibly specific to have much of an effect (e.g., adding a line like \u201cin the style of H. R. Giger\u201d), and even then, the impact wasn\u2019t obvious beyond the color pallet. My hopes for a story illustrated entirely by pixel art were quickly dashed.<\/p>\n Still, even when the scene illustrations aren\u2019t perfectly on-topic or realistic \u2014 think pirates with sausage-like fingers standing the middle of an ocean \u2014 there\u2019s something about them that give AI Dungeon\u2019s storylines greater weight. Perhaps it\u2019s the emotional impact of seeing characters \u2014 your<\/em> characters \u2014 brought to life in a sense, engaged in battling or bantering or whatever else makes its way into a prompt. Science has found as much.<\/p>\n What about the \u2014 ehem \u2014 less SFW<\/em> side of Stable Diffusion and AI Dungeon? Well, that\u2019s tough to say, because it\u2019s nonfunctional at the moment. When this reporter tested a decidedly NSFW prompt in AI Dungeon, the system returned an error message: \u201cSorry but this image request has been blocked by Stability.AI (the image model provider). We will allow 18+ NSFW image generation as soon as Stability enables us to control this ourselves.\u201d<\/p>\n Image Credits:<\/strong> AI Dungeon<\/p>\n<\/div>\n \u201c[The] API has always had the same NSFW classifier that the official open source release\/codebase has in the default install,\u201d Emad Mostaque, the CEO of Stability AI, told when contacted for clarification. \u201c[It] will be upgraded soon to a better one.\u201d<\/p>\n Terranova says that Latitude has plans to expand image generation with emerging AI systems, perhaps sidestepping these sorts of API-level restrictions.<\/p>\n With time, I think that\u2019s an exciting future \u2014 assuming that the quality improves and objectionable content doesn\u2019t become the norm on AI Dungeon. It previews a whole new category of game whose artwork is generated on the fly, tailored to adventures that players themselves dream up. Some game developers have already begun to experiment with this, using generative systems like Midjourney to spit out art for shooters and choose-your-own adventure games<\/a>.<\/p>\n But those are big ifs. If the past few months are any indication, content moderation will prove to be a challenge \u2014 as will solving the technical issues that continue to trip up systems like Stable Diffusion.<\/p>\n Another open question is whether players will be willing to stomach the cost of fully illustrated storylines. The $10 subscription tier nets 250 illustrations or so, which isn\u2019t much considering that some AI Dungeon stories can stretch on for pages and pages \u2014 and considering that crafty players could run the open source version of Stable Diffusion to generate artwork on their own machines.<\/p>\n In any case, Latitude is intent on charging full steam ahead. Time will tell whether that was wise.<\/p>\n<\/div>\n![\"AI](\"https:\/\/www.ghanamma.com\/gp\/wp-content\/uploads\/2022\/09\/Screenshot-24-e1662660569307.png\")
<\/p>\n![\"AI](\"https:\/\/www.ghanamma.com\/gp\/wp-content\/uploads\/2022\/09\/Screenshot-2022-09-13-133525.png\")
<\/p>\n