VRChat 2026.4.1 adds Safety Center as AI label decision draws criticism

VRChat went live with Release 2026.4.1 on October 28. According to the official Steam post, the patch introduces a Safety Center, improved face tracking and a new login flow, and the developer warns that it is not a network-compatible release.

In the same days, VRChat Inc. told its community on the official Ask Forum that it will not add an "AI-generated" label or AI filters to the platform, a decision that has drawn criticism from the press and from players.

What 2026.4.1 changes

The Safety Center is the most consequential addition. VRChat's open beta notes describe it as a screen that tells users whether their account is in good standing and, if not, what they did wrong and what their current punishment is. It sits behind a button at the top left of the Settings page.

Behind it is a new approach to rule breaks. In its 8 October Developer Update, the developer explained that an avatar-based violation used to mean an account suspension, while now it leads to a temporary restriction on using certain avatars, with the account still accessible. Players will also be able to appeal certain infractions from the client, including suspensions. The developer adds that it is still implementing the backend changes and rolling out the new policy slowly.

The login page was, in the developer's words, one of the oldest pieces of UI in the client. On PC, players can now log in by scanning a QR code with the VRChat mobile app, and anyone whose Steam account is not linked to a native VRChat account gets a one-time prompt to link it. The Steam post lists improved face tracking without further detail.

The AI decision

In "Where we're landing on AI in VRChat", VRChat says a "Human-authored" mark and a "Tools used" field are being considered but not scheduled. The developer argues that some creators would try to pass off generative or assisted work as their own, and that labels could encourage "sleuthing" and harassment of creators. According to the post, 21,595 people filled out its survey within a week, and 95% of respondents want some kind of label for AI usage, with little agreement beyond that.

Engadget led with exactly that gap: no labels despite 95 percent of respondents wanting one. Its verdict is that the post seems to want to "make people with many different attitudes toward AI all happy". On the Ask Forum, one player called it an "absolutely dog water, mealy-mouthed response". The issue has been building for weeks. Kotaku had reported a community developer calling for a boycott over a "rampant plague" of AI, with many of the most popular worlds displaying the protesters' banner in the app. In an earlier statement, VRChat had said it was updating its Creator Guidelines immediately and was "open to the idea of labeling".

On Steam, reviews are arriving far faster than usual: 1,667 new reviews in 48 hours, about 834 a day against 37 a day before. Overall, the game stands at 74% positive across 272,061 reviews, rated Mostly Positive. On October 30, Steam counted 36,755 players at once, with a peak of 54,073 that day, placing VRChat 36th on the most-played chart.

VRChat is free to play and still listed as Early Access on Steam. Returning players should note that 2026.4.1 is flagged as not network-compatible. Spookality 2026, with its creator-made horror-themed worlds and avatars, is still running: a DJ event takes place at Camp Whispering Screams on October 31 at 5PM PT, and visitors to that world on October 30 and 31 receive camp-themed hats. The winner rows for creator-made worlds and avatars stay up until November 2 at 2pm PT, the themed shop content until December 18.