https://www.independent.co.uk/news/world/americas/us-politics/russian-trolls-hilary-clinton-fake-news-election-democrat-mark-warner-intelligence-committee-a7657641.html

The link above is to the source article in question; it contains a video of Senator Mark Warner, chairman of the Senate Intelligence Committee. The quote which appears to have sparked the sensational headline is by him, as follows:

We know about the hacking, and selective leaks, but what really concerns me as a former tech guy is at least some reports and weve got to get to the bottom of this that there were upwards of a thousand internet trolls working out of a facility in Russia, in effect taking over a series of computers which are then called botnets, that can then generate news down to specific areas Its been reported to me, and weve got to find this out, whether they were able to affect specific areas in Wisconsin, Michigan, Pennsylvania, where you would not have been receiving off of whoever your vendor might have been, Trump versus Clinton, during the waning days of the election, but instead, Clinton is sick, or Clinton is taking money from whoever for some source fake news.

 

First, let me address Warners claim to be a former tech guy. The Wikipedia article of him writes that he was an investor in at least two mobile phone companies, and co-founded Capital Cellular Corporation. Him investing in and being the co-founder of mobile phone companies does not, in anyway, make him an authority figure on home computing or networking, the major fields at hand.

With that out of the way, let me explain what the botnets mentioned by Warner are. A botnet is any number of Internet-connected devices running what are known as bots. A bot is a program or application that runs repetitive, automated tasks, also known as scripts, over the Internet. Examples of potential tasks for bots would be spamming voting systems, DDoS (Distributed Denial of Service) attacks (essentially, loading a web page multiple times in an attempt to overload the server hosting it), leaving comments or sending messages en masse (spam), or, in the case of a non-malicious task used by search engines like Google, web crawling (the process of fetching and analysing URLs to index them and allow them to appear as search results).

Botnets, as a typically widely dispersed network of computers, are the means of distribution, often using their geographic differences to confuse the protocols in place to prevent spamming by a single IP address or geographic location. This can be incredibly effective at overloading web servers or posting fake comments.

Warning states that Russia has nearly a thousand Internet trolls in its possession; this is absolutely believable, as Russia has utilised a PR firm to leave comments in support of Russia in articles and videos related to the Ukrainian crisis, as I mentioned earlier. However, Warner goes on to state that not the trolls, but the botnets were the author of news articles:

there were upwards of a thousand internet trolls working out of a facility in Russia, in effect taking over a series of computers which are then called botnets, that can then generate news down to specific areas

This is nonsensical; I can only assume he misspoke, intending to state that the trolls authored the articles, and circulated them as comments or posts and distributed them en masse via botnets. As I stated earlier, bots are built for repetitive, typically simple tasks; they by definition are not advanced AI, and thus not able to generate coherent text in any language. Continuing from Warners quote, he goes on to state, somewhat clumsily, that he has been informed (it is not stated in the quote or article by whom) that they may have been targeting specific areas, in this case, swing states.

Its been reported to me, and weve got to find this out, whether they were able to affect specific areas in Wisconsin, Michigan, Pennsylvania, where you would not have been receiving off of whoever your vendor might have been, Trump versus Clinton, during the waning days of the election, but instead, Clinton is sick, or Clinton is taking money from whoever for some source fake news.

The way he phrases this statement makes it quite unclear to me exactly what he means; he does not discuss means of distribution (e.g., a social media website such as Facebook or video sharing platform such as YouTube), nor does he discuss how any geographical targeting may have taken place. Such targeting would not be possible through Facebook or YouTube unless the sites themselves were compromised; as I stated in our discussion, it is simply not possible unless a site is compromised to force posts to be seen by users or to otherwise tamper with the content delivered to them. In order to see posts on Facebook, one must be friends with or have liked the person/page that posted, or be friends with someone who has liked the post in question. The only two means of distribution that could be effectively distributed with a botnet and targeted to recipients based on geographic location could be SMS text messaging (the geographic location is within the area code of phone numbers) or email (the geographic location, unless obvious from the email server domain, would have to be obtained through other means, likely crawling).

I question Warners competency regarding the subject at hand. I was able to find another article, located here: http://www.cnbc.com/2017/05/19/inside-russias-social-media-war-on-america.html

It mirrors Warners quote as follows:

Democratic operatives searching for explanations for Clinton's loss after the election investigated social media trends in the three states that tipped the vote for Trump: Michigan, Wisconsin and Pennsylvania. In each they found what they believe is evidence that key swing voters were being drawn to fake news stories and anti-Clinton stories online. Google searches for the fake pedophilia story circulating under the hashtag #pizzagate, for example, were disproportionately higher in swing districts and not in districts likely to vote for Trump.

 

The statement about Google trends piqued my interest; if this were true, it would indeed be evidence of such targeting. However, if one looks at the Google trends results (which you can do here), youll find that the term what is pizzagate and #pizzagate both are much more popular on average in rural areas than in urban areas, far from the claims of disproportionate representation. Unless Im missing somehow missing something, this article is quite misleading. Of course, as Im not a Senator, Im not gifted enough to participate in the hearings, but I remain unimpressed by the evidence made available to me. If theres anything substantial about this that Ive missed, please let me know.