Roblox platform safety and user risks

The Norwegian Media Authority (Medietilsynet) issued a report titled ‘Roblox under radaren’, warning of significant risks to children and adolescents on the gaming platform. The analysis claims the platform’s design prioritizes profit over well-being, utilizing mechanisms that encourage addiction and impulsive spending through the virtual currency, Robux. Specific concerns include features that mimic gambling, such as loot boxes, and the difficulty parents face in monitoring user-generated content and interactions with strangers. While the platform offers creative opportunities, it presents risks including cyberbullying, cybergrooming, and exposure to inappropriate content due to inconsistent moderation. Experts noted that children can often bypass existing safety protections and communicate with adults without effective age verification. Recent research in Italy by YouGov for Kaspersky involving 525 children aged 10 to 16 further highlights these vulnerabilities. The study found that approximately 32% of surveyed youth reported contact with unknown individuals during online gaming sessions. Specifically, 8% of participants shared sensitive personal data with strangers, and 7% sent personal photographs or images. Investigations into Roblox, which hosts over 100 million daily active users, indicate that users may be directed from the gaming environment to external chat applications like Discord or Telegram to trade items or access content. Such transitions can expose minors to violent messages or coordinated online groups that celebrate real-world incidents. In October 2026, the platform faced allegations regarding the unauthorized collection of 145.5 GB of 3D models. Operators of Open3DLab, SFMLab, and SmutBase reported that an automated web crawler, suspected to be linked to Roblox, downloaded massive amounts of data over a 34-hour period. This collection is controversial as it includes 62.6 GB of adult-oriented content, which operators suggest may have been harvested to serve as training data for generative artificial intelligence models.