"There's something strange about search engines, it's like you found oil in a world that hasn't invented internal combustion engines yet. Raw materials too much to the point where no one knew how to benefit from them."
Dialogue from the 2014 film Ex Machina.
For more than two and a half decades, we have become accustomed to the sight of a blank search bar on a blank page, with the "Google" icon appearing in different colors, becoming our first gateway to the Internet. Over the same period, there have been many competitors who have tried to capture even a small percentage of the online search engine market, but none of them have been able to get close to that fortress.
But the walls of the castle began to shake recently, since the moment the chatbot "Chat GPT" appeared, and launched its exciting journey, everyone felt that something was about to change, so that the alarm began to sound within the walls of Google Castle itself just a month after the launch of the new robot. Google, perhaps for the first time since its founding, feared a massive technological change was looming that could disrupt its core business model, which is Internet search. (1)
We're not here to predict Google's sudden disappearance or collapse, in fact, the company is well managed and has really smart management. However, change is certain, and these massive changes present enormous challenges for Google search, perhaps the truly largest and first of its kind over the past two decades. When Google search is affected, our relationship with the Internet itself will be profoundly affected, which begs us to ask: Is it time to replace traditional search engines, which rank and rank web pages, in favor of AI-based language models that provide direct answers to our questions?
To answer this question, we will first need to ask several partial questions: Can the new models fully understand the complexities and contexts of human languages, so that we can trust their answers to our questions? More importantly, should we really go through such a massive change now? And what would the consequences be for a company the size of Google if traditional online search methods were to change? But what will the consequences be for the Internet system as we know it in its current form?
What's interesting here is that these questions don't have direct answers, and it won't be that simple.
New search experience
Google won't redesign its search homepage to look more like a chat GPT chat, but it will place the AI's answer to the question at the top of the current search results page. (Google)
Google's search engine works primarily with the support of artificial intelligence algorithms, which play their role in ranking the pages of sites in search results. (2) The dilemma now is that chatbots, based on generative AI models, offer us new promise to reimagine our relationship with the entire online search process. Just as Google made us this promise from its earliest beginnings, our gateway to the Internet may change with it.
This is what Google itself sees, and has been planning for a long time, to show the first manifestations of this new world at the recent Google I/O conference, in which it explained that the future of search on its engine will depend on artificial intelligence, which it called the "Search Generative Experience" feature. The company won't rely entirely on chatbots, nor will it redesign its search homepage to look more like a chat page with a GPT chatbot, but instead will place the AI's answer to the question at the top of the current search results page. (3)
When the user types a question in the Google search rectangle, the regular search results will appear in Google as usual, but above it will appear an orange rectangle, and after a few seconds a short answer to the question written by artificial intelligence will appear inside it. On the right, there are three links to sites that contain information confirming what is stated in this summary.
In the near future, the new look of the Google search results page will become, artificial intelligence first, and not blue links from several websites as we have been accustomed to over the past 25 years. Artificial intelligence powered by one of the most sophisticated large language models to date, the PaLM 2 model, as well as the unified multitasking model "MuM" that Google uses to understand multiple and different types of media.
An old change plan
Knowledge graph (Shutterstock)
Google has been on this path for a long time, constantly pushing for UI changes to provide answers to questions asked on the search page itself, rather than providing the answer in the traditional way from a website that appears in regular search results.
These changes included the Knowledge graph, a huge database of information that enables Google to provide instant and correct answers to some user questions, which appear when searching for specific facts, people or places, and this process occurs within the Google search page, without having to go or click on links to any other site. (4)
This is in addition to featured snippets, known as the result zero, which is a result that appears at the top of the search page, intended to provide the user with a direct answer to his question without having to click on the site link that provides that answer. (5) These changes push traditional search results away from the top of the search page, but they are often useful in getting a quick answer to a simple question, which is all you want to get from the search process in this case.
But why would this change in the way Google searches go online? In other words, why does the user decide to turn to other methods and places to search for information?
Why has Google become so bad?
Complaints about the deteriorating quality of Google search results have begun to gain some momentum over the past few years. (Shutterstock)
Google search has naturally led to a radical change in the Internet and in the way we access information and facts in the world around us. Its enormous control over our lives has moved it from a digital index, with algorithms combing various pages and websites, to become synonymous with the act of searching for any information that has a digital presence. Perhaps it has turned in itself into an immense entity of knowledge in our human imagination.
Despite all that control, complaints about the deteriorating quality of Google search results have begun to gain some momentum over the past few years. (6) Of course, there are several reasons for this, the most important of which is the deterioration of the quality of the content itself on the Internet in general, but even this reason Google search played a key role in it.
What Google promised at its inception was to organize valuable world information, but over the past quarter century, the opposite has happened; it organized a huge amount of world information to fit Google's search engine, or to reserve its place on the front page results of this giant index.
The first problem here is the large number of ads, which may occupy the first half of the search results page on Google, simply because advertising is the most profitable source of income for Alphabet, the parent company of Google, in 2022, its advertising revenue reached $ 224 billion, which represents about 80% of its total revenue during the same year. (7)
The second problem, created by the company itself, is the inflation of the SEO monster, this enormous monster, so to speak, is starting to eat away at the entire search user experience, which you will notice yourself when searching on Google for almost anything. As such, Google has reshaped the nature of content on websites; there is an endless race between blogs, news sites, content creators, and anyone who wants to sell a product, to make sure that their site's title ranks on the first search page of Google, turning it into something like a huge digital marketplace.
Over time, an entire economy based on this industry was born; it is difficult to determine the size of the economy based on Google search alone, but according to the economic impact report of the company's various services for the past year 2022, Google Search, Google Play Store for apps, YouTube, cloud services and advertising tools contributed to the production of more than $ 700 billion in economic activities for millions of companies, nonprofits, publishers, content creators and developers in America alone. (8)
The user has less confidence in the results displayed on the search page than in human experiences. (Shutterstock)
There are sites that only specialize in writing articles that target search engine technologies to occupy the first results by targeting keywords that the user is actually searching for, so there are functions that are entirely dedicated to understanding how Google search ranks pages, how search algorithms work, and other complex technical matters.
The problem is not that search techniques or algorithms have become worse, but rather in the fact that the user has become less confident in the results displayed by the search page compared to seeing human experiences in videos on YouTube or Tik Tok, or live discussions on social networking sites, the most famous of which at this point is the site "Reddit", which the user adds after asking any question in Google search, to show him the results from the site "Reddit" specifically and not from other sites that Top search results.
For example, when a user searches for reviews of a new smartphone, or any other product they intend to buy, it often actually happens that they are looking for those reviews on sites where they find people or users like them, or looking to watch a live experience of the device in YouTube videos, simply because they want to see what real people and enthusiasts are talking about buying that product, rather than reading the top 10 results on Google's first search page, which often What are paid reviews from the phone manufacturer, and advertise its advantages anyway, without touching on a real human experience.
Of course, those opinions on social media sites are personal, and they may be wrong opinions about the product based on the experience of this or that person, but in the end they remain more realistic discussions for the user, and better than the results that appear from websites on the Internet. Maybe with search engines powered by AI answers, this stops, and sites and blogs are returning to providing content that focuses more on quality, cares about the user and provides a real human experience!
Different search experience
The idea of answering our questions in conversation and interacting with us is a method that we have become accustomed to over the past decade thanks to various chat applications, such as WhatsApp, Messenger and others, which was confirmed by Sundar Pichai, CEO of Alphabet, in an interview with the Wall Street Journal last April, pointing out that this is the company's vision for the future of the Google search engine, by interacting with the user in a conversation to answer His questions. (9)
Looking at social media applications, we will find that they have really changed our relationship with Internet search, and there is a new generation looking for its information in Tik Tok, Instagram photos, or YouTube videos, as many believe that these means provide better visual human experiences than the results they may find on the Google search page as mentioned.
On the new Google search page, three cards will appear to the right of the robot's summarized answer, containing links from pages to different sites confirming its answer to the question. But if you get the answer, you'll ask yourself the logical question: Why do I click on those links that will lead me to a page on another website?
This criticism is not directed at Google's new search engine only, because the chatbot of the search engine "Bing" displays source links to a number of sites that confirm its answer to the question as well, but whether the user clicks on this link or not, and clicks on it at a useful rate for the site, remains vague so far. But if we consider that the goal behind an AI-powered search engine is simply to summarize the information and display the most important information in one answer, this will eliminate the possibility that you will need to read more about it, especially if you are looking for quick and specific information. This brings us to the point of influence on the advertising ecosystem and the sites that live on its gains.
The content industry and marketing economy may need to restructure its strategies to keep pace with this new change in consumer behavior, and not move to the site page after taking what information it wants from within the Google page, and searching for new ways to develop the quality of content, because this is what large linguistic models will rely on to summarize the answers later. But herein lies the problem, can content provided by artificial intelligence be trusted, especially if it is our only source of information?
Subtle effect
Google search is so useful and pervasive in every aspect of our lives right now that it's hard to notice its enormous impact on us. We can easily recognize the impact of certain tech products on our lives, such as smartphones or social media, yet it is different with Google search, and despite it being one of the most lucrative businesses of all, we can't really grasp the exact magnitude of its impact.
"Our Web: More Knowledge, Less Understanding in the Age of Big Data" by philosophy professor Michael Lynch. (Social Media)
In his book Our Internet: More Knowledge, Less Understanding in the Age of Big Data (10), philosophy professor Michael Lynch begins by recounting an intellectual experiment: imagine a society where smartphones are too small and can be implanted directly into a person's brain as a chip. When a person thinks of a fact, the slide will tell him the appropriate answer, and this certainly has nothing to do with the Egyptian film "Lambi 8G" if you are wondering.
The important thing is to imagine now that generations and generations live in this cultivated segment, and those generations grow up relying on it completely to know what they know, to the point that they forget how people learned in the past: by observation, careful research, and reliance on logic and critical thinking.
Then imagine this: in a flash an environmental catastrophe destroys the planet's electronic communications network, destroying all the chips implanted inside people's brains, as if the whole world had suddenly gone blind. There will be no way to prove the facts, and no one will know anything anymore, because they simply do not know how to know!
Lynch thinks we're very close to getting to that point, we can't see the facts, we can't gain knowledge, and he thinks we're already no longer able to agree on how we know. With normal online search methods and our availability of information, what would it be like if we replaced AI-based search engines with those segments in Lynch's intellectual experience?
Monopoly of truth
Obviously, humans trust AI information more than they trust their species, because we often think that AI systems are objective, and their answers don't have prejudice. (Shutterstock)
When we encounter an answer from a robot, or when we imagine that there is some kind of automation going on in the background to produce that answer, we tend to believe the machine easily, which was tried to prove by a study from a major Polish university in 2020, which stated that more than 85% of test participants ignored the behavior of the robot, even as it deliberately produces wrong answers, and considered it a reliable reference for making their decisions. (11)
Of course, we need more research and studies to better understand this new human phenomenon, but it is clear that humans trust AI information more than they trust the information of their species, because we often think that AI systems are objective, and their answers do not contain prejudice, but on the ground, and because these systems are trained from human data and conversations in the first place, they are vulnerable to human bias and other defects.
This problem can be easily mitigated if these products take the time to test and inspect before they are released as products for public use, but the current speed and frantic race between Google and Microsoft has accelerated the process of integrating AI products into most of their services, and of course the most important and lucrative service is Internet search.
The other problem is that these large language models are artificial intelligence systems that are trained to recognize patterns in a very large number of texts on the Internet, such as books, conversations, articles, etc., and then train more with human help to provide better conversation and answers to the user. But while the answers we get may seem convincing and even reliable, they can be completely wrong, which is called hallucination. These artificial hallucinations occur when the robot answers you with a confident but unjustified answer in the data it has been trained on, and the term has been derived from the psychological concept of hallucinations in the human mind because of the common traits between them. The problem with robot hallucinations is that the answer may seem convincing and ostensibly correct, but in reality it is wrong. (12)
New search engines
The search engine is supposed to act as an index to organize information, like any self-respecting search engine, and should not act as an encyclopedia that presents the information itself. (Shutterstock)
Now those answers will come from the search page itself, so the responsibility for proving them will lie with Google, not on the pages of the various websites that provide this information or answers, which you can of course verify their credibility as before.
Maybe this works for a few simple questions, like when Lionel Messi was born or where he plays now, and simple things like that. But complex questions or problems, such as looking for health and life issues, buying an expensive item, or looking for other people's advice and experiences on different matters, all require personal judgment based on reliable experiences, and one summarized answer from AI models often won't suffice.
Taking this new approach will often risk diminishing the essence of Google's success — we are not used to Google's power to provide direct answers to our questions, but to help us easily discover information ourselves as promised. A search engine is supposed to act as an index to organize information, like any self-respecting search engine, and should not act as an encyclopedia that provides information itself, so it should not affect its search capabilities in order to provide immediate results and answers.
These problems, although not clear now, will worsen when new search engines begin to spread, and with them our behavior in searching and obtaining information changes, because we often will not get traditional search results with links from several sites on the Internet, with which we can assess the quality of content and information, and search more for facts. Instead, AI will produce one brief answer that it presents as objective fact, with a few links to confirm that answer.
In the end, perhaps Google search will disappear in the old form we are used to, and perhaps the Internet search economy will disperse over more than one player in the field, and the user will move between new search engines with artificial intelligence, or TikTok and YouTube search engines. This future will certainly be exciting, and most importantly, this competition may contribute to the creation of more quality and reliable content, or at least so we hope. But what is certain is that the way we have been accustomed to searching over the past two decades will change once and for all, perhaps forever.
————————————————————————————
Sources:
1) A New Chat Bot Is a 'Code Red' for Google's Search Business
2) How search works (just the basics)
3) Supercharging Search with generative AI
4) Introducing the Knowledge Graph
5) How Google's featured snippets work
6) Is Google Search worse? Many Seem to Agree
7) Advertising revenue of Google from 2001 to 2022
8) Our economic impact in the U.S. in 2022
9) Google CEO Sundar Pichai Says Search to Include Chat AI
10) The Internet of Us: Knowing More and Understanding Less in the Age of Big Data
11) People Blindly Trust AI Even When It's Clearly Wrong, Study Shows
12) The War of Artificial Intelligence. Has Microsoft triumphed over Google?