Inappropriate content filter: the safety filter for image blocks in web results
The leak points to a Superroot library that filters inappropriate images in the image block of web results. Google documents SafeSearch for Search; that this twiddler implements it is our reading.
01What the schema and the leak show
The schema object is IUInappropriateFilterTwiddlerProto. IU stands for Images Universal, the block of images Google mixes into ordinary web results. RESONEO describes it as removing inappropriate images from that block.
The 2024 leak names a matching library. A field description in the data on images inside web results points to iu_inappropriate_filter_lib.h, a file under superroot/impls/images/quality/safesearch, for fuller filtering of IU images. The path places the filter in Superroot and in a SafeSearch folder. Unlike most twiddlers, a filter removes rather than reorders.
02What Google says about SafeSearch
Google's help page says SafeSearch tries to identify and filter inappropriate content, which includes nudity, explicit content, violence and gore. Users choose between Filter, which blocks explicit results, Blur, which blurs explicit images, and Off. Filter is the default when Google's systems indicate a user may be under 18.
For site owners, Google's image guide says to make sure Google understands the nature of a site so SafeSearch filters can be applied where appropriate. Neither page mentions a twiddler or the IU block by name.
03E-E-A-T and the system map
This is a safety filter, not a quality or trust signal. It removes images that are inappropriate for a setting, whatever the expertise of the page. On our map it belongs to the twiddler layer and to no E-E-A-T system. What a site owner can do is what Google asks: label adult content correctly so that it is filtered where it should be and not elsewhere.