Ever wonder at how Google Search arrives at those authoritative looking answers that often appear to queries, at the top of the first results page?

The fact is, Google is quickly developing and implementing new levels of AI powered mediation into its search engine platform.

But “featured snippets,” Google’s term for those top result answer boxes, is a small step in what’s to come.

The tech giant’s goal is to meld their search engine advertising empire with “meta-info” AI technology that will take technocratic internet info gatekeeping to a whole new level.

Currently many queries on Google return an authoritative looking first result.

The results can actually be quite non-authoritative, to say the least.

For example, ask Google search the question “how many humans can the earth sustain,” and Google presents this reply:

“An average middle-class American consumes 3.3 times the subsistence level of food and almost 250 times the subsistence level of clean water. So if everyone on Earth lived like a middle class American, then the planet might have a carrying capacity of around 2 billion.” 

The “answer” is a snippet from a webpage, in this case from the website.

Note the query to Google wasn’t “how many people could the earth sustain at current ‘middle class American’ living or consumption standards.”

The selected snippet reflects some pretty clear ideological biases.

Google has its own term for the AI selected answers at the top of many of its search engine query results: “featured snippets.”

Their explanation of how featured snippets are selected refers to their “automated systems” and algorithms:

“…Featured snippets come from web search listings. Google’s automated systems determine whether a page would make a good featured snippet to highlight for a specific search request. Your feedback helps us improve our search algorithms and the quality of your search results.”

Of course, Google gate-keeps those authoritative snippets, both via algorithmic training and manual interventions:

“…Our automated systems are designed not to show featured snippets that don’t follow our policies. However, since the scale of search is so large, we also rely on reports from our users. Your reports help us improve our search algorithms to avoid issues in the future.

“We manually remove any reported featured snippets if we find that they don’t follow our policies. If our review shows that a website has other featured snippets that don’t follow our policies or the site itself violates our Google Search Essentials, the site may no longer be eligible for featured snippets.”

Featured Snippets are a step beyond simply ranking website links that may offer relevant info related to user queries.

But Google has a much bigger transformation in store for its search engine platform.

Fom Search Engine to AI Expert: Google’s BERT and MUM

The Google search platform will likely soon see changes that make the tech giant not just a filtering gatekeeper of website info, but an AI-powered synthesizer of information.

Two of Google’s newest technologies, BERT and MUM, use AI natural language processing and neural net learning. 

BERT, which stands for Bidirectional Encoder Representations from Transformers, is a deep neural network AI system that can process natural language queries and provide results based on whole sentence understanding.

That “understanding” is based on natural language understanding (NLU), and sentiment analysis that puts every word in a search query in context, by relating the words to each other.

Continual machine learning on huge data sets, and neural net learning technology are all part of the AI “magic.”

Web page optimization experts have had to adjust to the BERT prioritizing and producing search results to queries, according to

The website quoted marketing expert Lemuel Park, CTO at BrightEdge, concerning BERT:

“BERT is Google’s next iteration in its long-running effort to better map search results to search intent. It is Google’s neural network-based method for natural language processing (NLP) to understand queries that are more conversational based in nature.

“This [is] an algorithm change and not an update. As a marketer this means increasing the specificity and depth of content and working farther into the longtail, or queries using more than three words.”

MUM, or Multitask Unified Model, is Google’s latest announced AI technology.

The company has called it “a new AI milestone for understanding information.”

A google engineer explained that MUM essentially turns Google search into an AI entity that can process queries and provide human-like answers, as opposed to mere website links or selected snippets.

In a blog post, the engineer offered an illustration of how MUM can work:

“You’ve hiked Mt. Adams. Now you want to hike Mt. Fuji next fall, and you want to know what to do differently to prepare. Today, Google could help you with this, but it would take many thoughtfully considered searches — you’d have to search for the elevation of each mountain, the average temperature in the fall, difficulty of the hiking trails, the right gear to use, and more. After a number of searches, you’d eventually be able to get the answer you need.

“But if you were talking to a hiking expert; you could ask one question — “what should I do differently to prepare?” You’d get a thoughtful answer that takes into account the nuances of your task at hand and guides you through the many things to consider…

“…Today’s search engines aren’t quite sophisticated enough to answer the way an expert would. But with a new technology called Multitask Unified Model, or MUM, we’re getting closer to helping you with these types of complex needs. So in the future, you’ll need fewer searches to get things done.” 

It’s true that most people using search engines want to find information efficiently. But MUM represents a new level of intermediary control by a tech company over what information gets through to users, and what doesn’t.

Most people using search engines are looking for third-party sources of information. They are not looking for information drawn from various sources and synthesized by a Google AI system, into a kind of “meta-information” result, given by a conversational AI entity / interface. 

With MUM, Google essentially relegates all websites into mere potential wells of information, from which its AI selects and assembles what it decides is relevant information.

As Google itself acknowledges, a search engine becomes an “expert” entity, giving an answer by drawing on sources it deems relevant.

If Google frankly offered MUM as part of a unique platform and interface, there’s no doubt people would flock to it, as they have to OpenAI’s ChatBot GT3.

But Google appears to want to integrate MUM while still retaining all that revenue from its search engine related advertising empire. Website owners pay billions every year with the expectation that their website info and listings will place high in Google and other search engine provider results.  

So Google may believe it cannot simply offer MUM as a unique product, undercutting its classic search engine platform.

Microsoft, which invested $1 billion into OpenAI in 2019, is now looking to invest an additional $10 billion, and they obviously see the relationship between current search engine technologies, and where the future is pointing.

Google does, as well, obviously.

The question is, what does that mean for the rest of us trying to find unbiased, independent (from tech company meta-info control and manipulation) sources of information? 

And It Gets Worse

If the prospect of having a Google AI expert standing in as a meta-info provider sounds unsettling, there’s more to consider, concerning the current explosion of AI.

“Expert AI” may not only be in control of synthesizing answers to search engine queries within the next few years. These expert AI interfaces may be drawing on information mostly compiled by—you guessed it—AI bots.

Nina Schick, founder of Tamang Ventures, an advisory firm focused on Generative AI, recently said on Yahoo Finance Live:

“I think we might reach 90% of online content generated by AI by 2025, so this technology is exponential,” she said. “I believe that the majority of digital content is going to start to be produced by AI. You see ChatGPT… but there are a whole plethora of other platforms and applications that are coming up.”

Schick believes that so-called generative AI, or AI software platforms that specialize in generating creative and other kinds of written or visual content, will soon be used on an pervasive scale.

We detailed the dangers involved in tech companies controlling the production of creative content via AI and Software-as-a-Service (SaaS) platforms. (See “YOU WILL OWN NO SOFTWARE AND BE HAPPY—PART ONE” and “YOU WILL OWN NO SOFTWARE AND BE HAPPY—PART TWO.”)

Between AI created content, and AI meta-info internet search mediation, it’s clear that the mega tech companies who control AI, will be controlling and monetizing information acquisition, creation and dissemination on an even more pervasive scale than they already do right now.

Skip to content