From Mountain View they have announced that Google will stop answering silly questions. This change in the search AI reflects a much brighter future for the company.
Google is the search engine par excellence. Although certain statistics ensure that young people prefer to use TikTok as a search engine before Google, the Mountain View service offers a much more complete experience. One of its most interesting features is that of the fragments, which will soon receive a pretty big facelift with the help of artificial intelligence. Of course, with the nuances that this entails.
Although you may have already taken advantage of them, you probably don’t know that this feature is presented under the name of snippets. What are they exactly? These are those results that appear first after a search , and that have an appearance closer to a widget than to a common result. Snippets expanded in 2018, allowing for much more information to be covered, and have been one of Google’s most useful features ever since.
However, it seems that these fragments have a problem. While their popularity has been growing, they have also started behaving erratically . That is, they offer results to questions whose premise is already false, generating confusion. Also, they can give somewhat simplistic answers, and this is where Google turns to its new artificial intelligence.
Those of Mountain View plan to renew the AI that allows Google to understand the searches of its users , and the subsequent results. On their official website , they explain that, although the fragments are quite efficient when it comes to giving specific data such as the height of a mountain or the weather and time of an area, the users’ doubts do not stop there.
To do this, they give an example of an expert in mountaineering issues in front of their AI. Google is able to offer a concise and factual answer about the size of a mountain. However, when asking questions like “what should I consider to prepare before uploading?”, the expert would be able to offer much more relevant and accurate information .
Now, the service wants to raise its levels of sophistication. To do this, they will rely on a technology they have called the Multitask Unified Model, and they hope it will help them answer complex questions from users .
With this AI, Google would be able to better dissect your questions . Thus, they could know what you mean, and respond based on a cross-reference system that allows them to offer high-quality results. Of course, the above is still a promise.
Yes, Google has decided to stop answering questions whose premise is false. What do you mean exactly? Just imagine that you have typed in the search engine “When did Snoopy assassinate Abraham Lincoln?”. Automatically, Google will show you a detailed result about the day and place of death of the American figure. However, does it make sense to give the answer if Snoopy was not the one who did the deed?
For this reason, to avoid misunderstandings, Google will stop showing answers to questions that are poorly posed , and will start progressively.
Also, the hints and warnings system will see an improvement , offering more relevant results when users pose questions incorrectly. This way, Google will let you know that it has an answer that it fully trusts, but it might be slightly different from what you asked in the first place.
Now, we know that there are multiple cases of artificial intelligence being wasted when it is applied to internet platform algorithms. Without a doubt, it is worrying that in Google searches we will see an inadequate functioning as it happens, for example, in Instagram and Facebook .
With Google’s AI intervening in the search engine, the following questions are automatically generated: will the search results be entirely subject to our request? what percentage depends on the interpretation of artificial intelligence? When it comes to content recommendation, this technology is not always the best solution.