VRChat Bans Misleading Thumbnails, Including AI-Made, In Guidelines Update
VRChat's own statement names the bottleneck as discovery rather than generation, which is why the near-term fixes target world ranking and feedback instead of an AI tag it cannot reliably verify.
Key Facts
- VRChat Inc. updated its Creator Guidelines on October 7 to prohibit thumbnails that do not match the actual world content, with the rule applying whether or not AI generated the image.
- VRChat says users can use AI for translation, grammar checking, coding assistance, and personal tools, while generating entire worlds, assets, or Udon systems remains under discussion.
- The user survey on where to draw the line for AI content on VRChat runs for seven days starting October 7.
- A community group tied to the NO AI thumbnail movement has more than 6,000 members.
- VRChat plans experimental content discoverability measures within the next few weeks and will publish progress and future plans in about six weeks.
Reporting from 5 sources: Automaton, Denfaminicogamer, Inside, GameBusiness.jp, and 1 more.
VRChat Inc. published its position on AI use and updated its Creator Guidelines on October 7, prohibiting thumbnail images that do not match the actual content of a world. The rule applies whether or not the image was generated with AI. World names, tags, and descriptions are also covered. The company framed misleading presentation as a problem that predates generative AI, and said it undermines the person-to-person connections the platform values. VRChat is free to play on Steam, Viveport, Meta Quest, Pico, iOS, and Android, and does not require a VR headset. The update followed pressure from creators B1G and Patroll, who called for a stated position; a related community group has more than 6,000 members. A user survey on where to draw the line for AI content runs for seven days from October 7. Meanwhile, labeling AI content remains undecided.
"Presentations that deceive viewers are not appropriate for VRChat," the company wrote in the announcement. The updated guidance is paired with an infographic sorting acceptable AI uses from contested ones.
- Translation, grammar checking, and coding assistance: acceptable.
- Smart tools such as Unity and Blender: acceptable.
- Generating entire Udon systems, assets, or whole worlds: still under discussion.
The company also said AI makes it possible to publish more worlds in a shorter period than before, and that worlds which took months or years to build may become harder to find. Its planned fixes target discovery: improving how worlds surface, lowering the display ranking of mass-produced low-quality content, and adding a way for users to send simple feedback on worlds. Trial measures are due within the next few weeks, with progress and future plans to follow in about six weeks.
On an "AI" tag and filters, VRChat said it will weigh the trade-offs. Tagging has benefits, but drawing the line on how much AI use requires a tag is difficult, and only honest creators would use one. It added that distinguishing AI-generated content is getting harder.
Synthesized by Yomimono from the 5 cited sources below, including Japanese-language reporting where cited, then editorially reviewed before publishing.
Sources
- Automaton 『VRChat』公式、クリエイターのAI利用について声明発表。ワールドの粗製濫造や、サムネイル詐欺への懸念を受けて
- Denfaminicogamer 『VRChat』AI生成を含む「誤解を招くサムネイル」を全面的に禁止へ。ガイドラインを更新し、AI利用に対する見解を示す
- Inside VRChat、AI利用に関する見解を発表―サムネイル規定を更新、AI生成の扱いは議論継続
- GameBusiness.jp VRChat、AI利用に関する見解を発表―サムネイル規定を更新、AI生成の扱いは議論継続
- PANORA VRChat公式がクリエイターのAI利用について見解を発表 実際と異なるサムネイルの使用を禁止に