Google clarifies how search autocomplete works
After questions were raised about it possibly being politically motivated.
In yet another odd development of the 2016 campaign season, Google's autocomplete feature -- not the actual search results, just autocomplete -- has come under scrutiny of potential bias. A YouTube video posted by SourceFed (with a followup, posted here) picked up attention after claiming that Google failed to link "Hillary Clinton" with "indictment" in its autocomplete, unlike competitors Bing and Yahoo (that Yahoo is powered by both Bing and Google seems to have gone unmentioned), even if you spot it the i-n-d. So, clearly Google is in the tank for Clinton, right? As it turns out, an alternate and more accurate explanation becomes clear with some knowledge of how Google's algorithms work.
Contrary to the way SourceFed describes the inner workings of a search engine, Google mentions that its autocomplete does filter for terms, particularly ones that could link someone's name with things that are potentially hurtful or disparaging. This isn't especially new information, as Google has previously manually adjusted its results for copyright reasons, and has even been fined for defamatory results.
Searching 'Donald Trump law' suggests lawn, but not his numerous lawsuits.
If you're still not sure it works that way, one could try a similar test with Clinton's presumptive competitor in the presidential race. While Donald Trump has had stories written about him that include the word "rape" or "lawsuits," searching his name plus ra- or la- shows neither word as a suggestion. A blog post by marketer and SEO expert Rhea Drysdale goes into more detail about how this works and why, for those interested. Finally, simply typing in Hillary shows a top autocomplete suggestion with "Clinton email" for reasons that should be obvious by now. That single result can't conclusively prove Google isn't hiding something, but if it is, then it's doing a terrible job of warning people off of the ongoing investigation.
