There are many language Wikipedias about debunked race models that either say nothing that stops them being mistaken for current science, or where the qualification about the racist content is not early in the text. This is in English as much as elsewhere, for example wikt.
As a consequence of Wikimedia projects having no systems-based way to tag or flag scientific racism content (like articles in multiple languages about "Negroid race" or "Mongoloid race") that might be picked up systematically by external search databases, public search services like Amazon Alexa and Google Search, are returning "Wikipedia" branded results that promote scientific racism with no qualification that the terms are outdated or debunked and not considered "science" or based on any biological fact.
This should be a WMF development project in line with the WMF strategy, at least for initial analysis of what can be done to tackle this problem, so raising as a Task to aid discussion.
Though the WMF avoid any censorship of content, correctly tagging "historic" concepts of race so that commercial databases can de-rank or correctly present those results as historic is not censorship, but the correct presentation of potentially damaging anti-educational material so that readers are not mislead by misinformation.