So, while his conclusions might be disingenuous, his assertions (manipulation) can be independently confirmed.
>Is there evidence of actual favoritism in Google’s search engine? Well, the European Union certainly thinks so, having fined Google $2.7 billion last year for having biased search results.
The fine was for an anti-trust ruling, where it was determined Google was promoting their own shopping results (showing them at the top of the page) over organic shopping sites, stifling competition. Favoritism for their own service, maybe, but nothing to do with pushing or manipulating results.
So that's why I suspected he had an agenda to push and was willing to spin the truth to argue for it.
You can argue that study [4] showed Google tended to have more Clinton-favorable results, but I don't personally think that shows manual manipulation. It's worth a critical read--in Issue 7, he seriously alleges that Google somehow realized this study was being conducted halfway through, and started serving unbiased results to subjects with Gmail accounts. If you accept that conclusion, okay, but I'd tend to believe that the results aren't as statistically significant as claimed.
Here's a read of some of the things they do out of legal obligation and other they do for selfish reasons and others for other reasons they believe are worthwhile: https://en.wikipedia.org/wiki/Censorship_by_Google#Search_su...
compare an autocomplete search for sarah joeng, robert muller, or roger stone on google, ddg and bing and look at SSE results. Bing DDG show relevant political issues, Google is coy about those.
Twitter and FBs issue is reliance and optimizing for engagement which leads to clicking and surfacing the most outrageous stuff at the expense of the ordinary.
Google search doesn’t have “an gagement problem”. They have black hat SEO problem, which is different, but they go beyond that and tweak some search results for political reasons. Not talking about terrorism or supremacy groups at all, but somewhat controversial social issues, (reproductive) sometimes tame issues (searching for candidates)[1]. The EU has fined them for “re-ranking” SERs and SSE effects.
>These data showed that Google’s search results favored Hillary Clinton (whom I supported) in all 10 positions on the first page of search results — enough, perhaps, to have shifted two or three million votes in her direction over time.<
Given that, 2-3MM vote swing (vs, what the outside estimate was under 50k votes for Russian bot sway), I’d say Google is the real problem in affecting democratic elections.
[1]https://www.usatoday.com/story/opinion/2018/09/13/google-big...