Roblox Downloaded 145GB of Models From Open3DLab Without Asking
On this page
On 27 October 2026, Open3DLab reported that Roblox had downloaded roughly 145GB of 3D models from its sites — including adult content — without asking permission or crediting the source. The incident centers on large-scale automated scraping of model repositories for use in Roblox’s platform.
The report raises familiar questions about web scraping, robots.txt, and content licensing in the age of AI and platform-scale data collection. Whether the material was used for training, asset ingestion, or another purpose was not fully clear from the report, but the volume and the inclusion of adult content made the episode notable.
Why It Matters
This is a vivid instance of the broader “data provenance” problem: automated systems harvest the open web at industrial scale, often ignoring norms of consent and attribution. It connects to copyright and fair-use debates, to how platforms build asset libraries, and to public trust in how companies obtain data. See AI Law and Justice, Culture, Communication, and Epistemic Trust, Model Collapse and Synthetic-Data Risk, and Open-Source and Open-Weight Models.
Sources
- Open3DLab. Roblox downloaded 145GB models from us (including adult content) without asking (news, 27 Oct 2026). https://open3dlab.com/news/roblox-downloaded-145-gb-of-models-from-our-sites/
- Hacker News discussion. https://news.ycombinator.com/item?id=50256985
Back to Recent Topics.