Stop showing Hitler on predictive and suggestive features exactly like you do not show gambling or porn or any other topic that demands someone to specifically search for it.
I found the same:
There is lack of transparency in the policy choices around both autocomplete and safesearch interventions. Here is a different angle on that problem in safesearch: The “SEARCHING FOR SOMETHING AND GETTING WHAT APPEARS TO BE EVERYTHING” section in the chapter on “TO REMOVE OR TO FILTER” in Gillespie (2018, pp. 186–194). An excerpt, from p. 187:
To be safe, search engines treat all users as not wanting accidental porn: better to make the user looking for porn refine her query and ask again, than to deliver porn to users who did not expect or wish to receive it. Reasonable, perhaps; but the intervention is a hidden one, and in fact runs counter to my stated preferences. Search engines not only guess the meaning and intent of my search, based on the little they know about me, they must in fact defy the one bit of agency I have asserted here, in turning safesearch off.[41]
- Jemima Kiss, “YouTube Looking at Standalone ‘SafeTube’ Site for Families,”Guardian, May 29, 2009, https://www.theguardian.com/media/pda/2009/may/29/youtube-google.
References
Gillespie, T. (2018). Custodians of the internet: Platforms, content moderation, and the hidden decisions that shape social media. Yale University Press. https://yalebooks.yale.edu/book/9780300261431/custodians-of-the-internet/ [gillespie2018custodians]
Without excusing the hidden interventions, note how “the one bit of agency” line both reflects and reconstructs articulations of search and perceptions of affordance.