True or False: Google Fakes Results for Social Engineering
September 13, 2016
Here in Harrod’s Creek, we love the Alphabet Google thing. When we read anti Google articles, we are baffled. Why don’t these articles love and respect the GOOG as we do? A case in point is “How Google’s Search Engines Use Faked Results for Social Engineering.” The loaded words “faked results” and “social engineering” put us on our guard.
What is the angle the write up pursues? Let’s look.
I highlighted this passage as a way get my intellectual toe in the murky water:
Google published an “overview” of how SEO works, but in a nutshell, Google searches for the freshest, most authoritative, easiest-to-display (desktop/laptop and mobile) content to serve its search engine users. It crawls, caches (grabs) content, calculates the speed of download, looks at textual content, counts words to find relevance, and compares how it looks on different sized devices. It not only analyzes what other sites link to it, but counts the number of these links and then determines their quality, meaning the degree to which the links in those sites are considered authoritative. Further, there are algorithms in place that block the listing of “spammy” sites, although, spam would not be relevant here. And recently, they have claimed to boost sites using HTTPS to promote security and privacy (fox henhouse?).
I am not sure about the “fox hen house” reference because fox is a popular burgoo addition. As a result the critters are few and far between. Too bad. They are tasty and their tails make nifty additions to cold weather parkas.
The author of the write up is not happy with how Google responds to a query for “Jihad.” I learned:
Google’s search results give pride of place to IslamicSupremeCouncil.org. The problem, according to the write up, is that this site is not a big hitter in the Jihad content space.
The article points out that Google does not return the search results the person running the test queries expected. The article points out:
When someone in the US, perhaps wanting to educate themselves on the subject, searches for “Jihad” and sees the Islamic Supreme Council as the top-ranked site, the perception is that this is the global, unbiased and authoritative view. If they click on that first, seemingly most popular link, their perception of Jihad will be skewed by the beliefs and doctrine of this peaceful group of people. These people who merely dabble on the edge of Islamic doctrine. These people who are themselves repeatedly targeted for their beliefs that are contrary to those of the majority of Muslims. These people who do not even come close to being any sort of credible or realistic representation of the larger and more prevalent subscribers (nay soldiers) of the “Lesser Jihad” (again, the violent kind).
My thought is that the results I expect from any ad supported, publicly accessible search system are rarely what I expect. The more I know about a particular subject—how legacy search system marketing distorts what the systems can actually do—the more disappointed I am with the search results.
I don’t think Google is intentionally distorting search results. Certain topics just don’t match up to the Google algorithms. Google is pretty good at sports, pizza, and the Housewives of Beverly Hills. Google is not particularly good with fine grained distinctions in certain topic spaces.
If the information presented by, for instance, the Railway Retirement Board is not searched, the Google system does its best to find a way to sell an ad against a topic or word. In short, Google does better with certain popular subjects which generate ad revenue.
Legacy enterprise search systems like STAIRS III are not going to be easy to search. Nailing down the names of the programmers in Germany who worked on the system and how the STAIRS III system influenced BRS Search is a tough slog with the really keen Google system.
If I attribute Google’s indifference to information about STAIRS III to a master scheme put in place by Messrs. Brin and Page, I would be giving them a heck of a lot of credit for micro managing how content is indexed.
The social engineering angle is more difficult for me to understand. I don’t think Google is biased against mainframe search systems which are 50 years old. The content, the traffic, and the ad focus pretty much guarantee that STAIRS III is presented in a good enough way.
The problem, therefore, is that Google’s whiz kid technology is increasingly good enough. That means average or maybe a D plus. The yardstick is neither precision nor recall. At Google, revenue counts.
Baidu, Bing, Silobreaker, Qwant, and Yandex, among other search systems, have similar challenges. But each system is tending to the “good enough” norm. Presenting any subject in a way which makes a subject matter expert happy is not what these systems are tuned to do.
Here in Harrod’s Creek, we recognize that multiple queries across multiple systems are a good first step in research. Then there is the task of identifying individuals with particular expertise and trying to speak with them or at least read what they have written. Finally, there is the slog through the dead tree world.
Expecting Google or any free search engine to perform sophisticated knowledge centric research is okay. We prefer the old fashioned approach to research. That’s why Beyond Search documents some of the more interesting approaches revealed in the world of online analysis.
I like the notion of social engineering, particularly the Augmentext approach. But Google is more interested in money and itself than many search topics which are not represented in a way which I would like. Does Google hate me? Nah, Google doesn’t know I exist. Does Google discriminate against STAIRS III? Nah, of Google’s 65,000 employees probably fewer than 50 know what STAIRS III is? Do Googlers understand revenue? Yep, pretty much.
Stephen E Arnold, September 13, 2016
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
