Harassment and personal space invasions have been a core privacy threat in social virtual worlds from the outset, often causing lasting harm to victims and reducing their willingness to engage with these technologies. Media coverage of such incidents further amplifies fear and discourages potential users. In response, platforms have increasingly emphasized proximity-control features that allow users to protect themselves both preventively and reactively – most prominently through safety bubbles and avatar blocking. However, no study to date has examined whether these features influence adoption-relevant factors. To address this gap, we conducted a video vignette-based online survey with 195 participants in Germany, each exposed to a mild personal space invasion scenario under one of three conditions: no safety features, safety bubble, or avatar blocking. Results show that both safety bubbles and avatar blocking positively influence relevant adoption factors such as trust, controllability, satisfaction, and use intention. Safety bubbles additionally improved perceived safety and reliability, appearing to be the more effective solution due to their preventive nature. Finally, implications of the findings are discussed.

Highlights• We actively trigger people’s tendency to treat robots as human-like social actors• Two sociopsychological mechanisms are applied in an industrial use case• Perspective-taking fosters cognitive trust in cobots• Anthropomorphism increases emotional distrust of participants• Human-like cobot designs may prevent collaboration instead of fostering it