Yesterday the Google I/O event took place in which the search engine company presented your latest news on artificial intelligence applied to its search and work services in the cloud. In The Verge they tell how Liz Reid, vice president of Google Search, showed them one of these novelties live.
Pay attention, because the scene has more relevance than it seems. Reid opens his laptop, logs into Google, and begins typing the following query into the search box: “Why is sourdough bread still so popular?” Then hit Enter.
As expected, the ‘traditional’ search results page loads almost immediately. But with a significant difference: above them, appears a blinking colored rectangle with the following legend: “Generative AI is experimental”.
And just a few seconds later, that rectangle is replaced by an artificial intelligence-generated summary, similar to this:
Thus, suddenly, the preferential place of the page is occupied by a few paragraphs that detail the advantages of the flavor of the sourdough and its prebiotic capacities. On one side, three links to websites whose content supposedly “corroborates” what was narrated in the summary.
Did you realize what just happened? Google is loading the web search in favor of the search for specific data… which does not require you to access any other website other than Google itself.
But, beware, this has not happened because Google is betting on AI, far from it… but because It’s just the culmination of a long-standing trend. in the Mountain View company.
‘Sgroogled.com’: when MICROSOFT launched ANTI-GOOGLE ads
From “we want you to leave Google as soon as possible” to semantic searches
Let’s go back almost two decades, to 2004. Larry Page, co-founder of Google, explained what his search engine was for: “We want to get you out of Google and into the right place as quickly as possible”. We talked about the times when Google used that famous ‘Don’t be evil’. Oh, the newspaper library.
Fast forward in time to 2012, the moment when Google announces something called the ‘Knowledge Graph’, its premiere in the field of semantic searches. Look how we described it at that time:
“Google now understands entities. If you search for ‘Marie Curie’, it will not only show you the most interesting results on the Internet with that string. It will also show you a short biography and interesting facts next to it, as you can see in the photo. In this case, being a scientist, you will see what discoveries she has made and where she has studied.”
All that extra information depended on the context: it could show sentimental partners (if we were looking for a famous person), books (if we were looking for a writer), etc. Of course, by now this is more than integrated into our idea of what a Google search is: is ‘what is expected’.
But think of how many times this mere information, available in a prominent place on the first page of results, It has been enough for you to verify some information and has dissuaded you from continuing to browse. Since then, this trend has only increased.
Do you remember when ‘surfing the Internet’ meant going ‘from island to island’ and not having everything at hand in the port?
Thus, sections such as “other questions from users” already offer Featured snippets of information, taken from third-party websites, directly on the Google results page, without the need for the user to leave the search engine at any time. And part of the links that are presented offer access only to more specific searches… in Google itself.
On other occasions, the first thing that appears under the sponsored links is an interactive space that, despite offering links to the competition, offers us the service we are looking for more directly through the Google website itself… such as the possibility of booking a plane ticket to Tenerife:
Google calls it ‘AI Snapshots’, and (for now) it’s optional
Returning to the present, to the AI-era results page, it’s fair to note that the summary generated by your language model will not appear for now when you search on Google: you must expressly enable a feature called ‘Generative Search Experience’.
Like the Knowledge Graph, it will not show up in every search we do, only when Google deems it useful. But, like the Knowledge Graph, it too is meant to stop being a mere ‘beta feature’ to happen to become, at some point, the new search. In ‘what is expected’ when accessing Google.
Image | Montage based on a scene from ‘The Lord of the Rings’
In Genbeta | What happened to the Google Toolbar… and the rest of the clones that for years cluttered the screen of our browsers