Developers of the social virtual reality platform VRChat are facing significant backlash from their user base after announcing they will not implement labels to distinguish AI-generated content. This decision comes despite a survey indicating that approximately 95% of users desired some form of AI content categorization.
VRChat serves as a popular hub for users to interact, create avatars, and develop games. The platform has seen a recent surge in AI-generated content, prompting a strong user demand for clear distinctions between human-created and AI-assisted creations.
A survey conducted by VRChat developers revealed that "95% of survey respondents want some kind of label for AI usage." However, the developers have opted against introducing such labels or filters. In an official blog post, they stated their reasoning, suggesting that while some creators would willingly label their work, others might misrepresent AI-assisted content as entirely hand-authored. The developers also expressed concern that such labeling could foster "sleuthing" behavior, leading to harassment of creators, even those not using AI tools.
Further complicating the decision, the survey showed that users proposed "3,141 different combinations of labeling," with no single combination receiving support from more than 10% of respondents. The development team concluded that they need to "understand this a layer deeper before making any kneejerk decisions."
This explanation has not appeased the VRChat community. Users have voiced their strong disapproval on social media platforms. Many question the purpose of conducting a survey if its results are not to be influential in decision-making. Others argue that the overwhelming 95% consensus on the need for filters should be a decisive factor, emphasizing that filters are intended precisely to allow users to avoid specific types of content.
As of this report, VRChat developers have not issued a further statement in response to the community's outcry, and the platform is continuing with its planned events, leaving many users frustrated by the perceived disregard for their input.