Google search enginehas an algorithm that does not understand the context
The search engine from Google is returning many times in its results a list of questions and answers relevant to the searched query, which I find very helpful if is done right.
The intention is to help the user get fast the searched answer without the need of reading an entire page of an article.
Google has a very complex algorithm when is about indexing the content of the website and the content. To get a good position in the search list, an article, or a site, must have very well optimised content for the search engines, what we know as being SEO.
The position of a page depends on the quality of the content, the use of: bold, italic, underline, headlines, with a balanced use of the focused keywords, optimised meta title and description, internal and external links, the page ranking, secure link (SSL – https) and a lot more.
Artificial intelligence is used in search engines
Artificial intelligence which is basically the algorithm Google is using to return results when a user is searching for something it is very advanced in these days being the result of many years of research and work of huge teams, plus users behaviour and feedback, but no matter how advance and powerful is, Artificial intelligence still can’t give or produce a response as the human brain can.
On the question/answers section of the Google search results page, Google is extracting from the articles what it seems to be the most relevant – the part of the text that is considered as being he best answer.
The answer is based most of the time on bold texts and is only considered as a true answer.
Apparently, the search engine can’t return results for a question that is asking for a solution by excluding things. Like “How not to lose weight” which will always return only articles about How to lose weight, or “Hair Clipper without batteries” which will always return results about Hair clipper with batteries. More about that in this article.
Not only once I have been in the situation of searching for some information and Google was unable to provide a helpful answer. I will show you in images what I mean.
If you ever did content optimisation on an article, you probably know that for that article to get a good position in the search engines, it must contain the focus keywords used in text bold, headlines (like h1 and h2), underline and italic text, title, first and last paragraph, have a specific density and so on. All these characteristics are helping the search engine to do the best judgement on what are your focused keywords and that you are focusing in convincing not only the reader but also the search engine which are those. For this reason is not enough only to list a number of keywords in a meta tag. The content must prove that those keywords are also properly used in the article by following SEO strict rules and into a logic context.
The context analysing seems to be the missing bit in how Google is picking the answers for the list of questions and for the moment only by reading more than google is considering as an answer, we can as humans, understand the the answer is wrong.
I asked recently Google if milk is good for gastritis and one of its answers as it can be seen in the image bellow it says that milk stimulated acid production for a question that will only accept a positive answer “Why does milk help gastritis”. But an increased production of acid is not helping with gastritis.
Not only that the answer was not carefully considered/cherry-picked from the full context of the content but the question that was formulated was built in such a way that it will only accept a positive argument and not a contradiction.
And this without even considering that maybe from a medical point of view the question it is a completely false information by already confirming that milt is helping on patients with gastritis and the Google answer is only underlying why this is true (which is not even true).
One more example
A simple search for a energy provider, Utility warehouse in this case is returning on Google results page a question that is looking for a confirmation if Utility warehouse it is actually cheaper that other providers.
The picked-up answer it is positive but is starting with a negation which again is proving that the search engine doesn’t see the entire context of a text and is only understanding the words as a value. For the search engine algorithm an answer to a question does not have to be logic. It only needs to fit the question.
The page that has the bold text maybe confirms from one end to another that actually Utility warehouse is cheaper than other providers, but the search engine considered important only the piece of text that was selected as an answer, which is making it yet superficial even if is very complex.
The reason the search engine is superficial even if is a lot f logic implemented in the back is because computer can’t still equalize human thinking and judgement.
A robot can do or know only what it has been told to do or know even if in the last years more and more information is proving that artificial intelligence exist. It only exist because of the data and the conditions that are introduced in a specific computer, call it robot.
As humans, a computer or robot can have reactions and provide responses based on the knowledge it has, but what robots don’t have yet is rationality and logic that is not based on knowledge but on instinct.
Any reasoning and logic a computer can have is still based on information, patterns, conditions and factors that needs to be defined and introduces by a programmer first into a system and then to be used by the system of a computer or machine. This is the reason a search engine so big as Google can’t return answers that people can trust, to its own question.
Another huge factor for false information on the search engines is the big increase of people involved in adding the content. If in the past the content was more introduced by professionals working in a certain domain and the subject was related to that domain, in our days people with no califications are writing articles based on the research they do ad not on experience and studies, for the only reason of making money from publicity from platforms like AdSense, Amazon Associates and many more.
Quality of content is decreasing because of monetisation.