Google still dominates search, but AI Overviews, a trailing Gemini and a quiet retreat from "Don't be evil" are testing users' patience. Could bad UX be what finally loosens Google's grip?
We write a lot about Google here, and just as much about UX. Recently I argued that user experience is the most important part of a brand. If that is true, it leads to an uncomfortable question for the biggest brand on the web: what happens when Google stops putting the user first? Could UX be what finally weakens Google’s monopoly?
Some of what follows is documented fact that anyone can check. Some of it is my own reading, as someone who has worked in UX for a long time. I will be clear about which is which.
What happened to Google’s “Don’t be evil”?
For everyone who grew up professionally in UX, Google was one of the most admired companies in the world and a role model. Its UX team was among the largest and most influential anywhere, and its motto, “Don’t be evil”, worked like a lighthouse for the whole discipline. That short, plain sentence framed core ethical principles: usability, accessibility, no dark patterns, transparency, not lying to users.
In 2015, when Google was restructured under Alphabet, the new parent company adopted a different line in its code of conduct: employees should “do the right thing, follow the law, act honorably, and treat each other with respect.” Google kept its own motto a little longer. In 2018 it removed “Don’t be evil” from the preface of its code of conduct, and the phrase survived only as a reminder in the very last line.
| Metric | Value | Change | What it means |
|---|---|---|---|
| 2000s | Don't be evil | Original | Google’s informal motto, later written into the 2004 IPO founders’ letter. Short, binary, hard to argue around. |
| 2015 | Do the right thing | Alphabet | Alphabet’s code of conduct adopts a broader, vaguer standard. Google keeps its own motto for now. |
| 2018 | Moved to the last line | Demoted | Removed from the preface of Google’s code of conduct. It survives only as a closing reminder. |
| Today | You can make money without doing evil | 1 of 10 | The closest echo of the original, now one principle on a list of ten. |
Read the newer wording closely and the questions write themselves. The right thing according to whom, and by what standard? Follow which law? (If Google has plenty of anything, it is lawsuits.) Act honorably under what framework? Treat each other with respect, but where exactly is the line? Nearly every legal fight Google has been in comes down, one way or another, to not respecting users or the laws of the countries it operates in.
The old sentence gave us a binary frame: whatever we do, let’s not be evil. And those of us who work in UX know very well how to be evil, precisely because we spent our careers removing that evil from interfaces. Today Google lists ten principles on its page Ten things we know to be true. The closest thing to the original is number six: “You can make money without doing evil.”
Is Google losing users’ trust?
This is where I move into hypothesis territory. My view is that Google had, and still has, the position to do an enormous amount of good, and the technology to amplify it. Instead, on data security, privacy and even cost, its relationship with users has gone dramatically backwards, and every change seems to make it worse.
I am not naive. I know stakeholders care very little about a UX paradigm and mostly want to watch their money grow. But how much money is enough? For what? How many lifetimes would it take to spend trillions of dollars? Questions nobody is going to answer.
What this race for revenue at any cost does is leave users more and more dissatisfied, and the cause is precisely the user experience. For the first time in a long while, Google finds other companies competing on ground it considered its own, especially in AI.
Gemini vs ChatGPT: why is Google third in AI adoption?
You would expect Google, with all the information on the planet, its quantum computing work, its processing power and a thousand other advantages, to be the clear winner of the AI race. It has been doing serious AI research since at least 2011, when Google Brain was formed, and it acquired DeepMind in 2014. It does not lack experience, technology or scientists.
And yet, going by the companies’ own reported numbers, Gemini is still third in adoption, behind ChatGPT and Meta AI.
These are self-reported figures, measured differently by each company, so read them as a rough ranking rather than an exact scoreboard. All three models get things wrong, like every other model. In my experience Meta AI is easily the weakest and the most prone to hallucinations, but it is built natively into products billions of people already use, which goes a long way toward explaining its numbers.
The real difference between ChatGPT and Gemini, though, is not the model. It is the experience. If I ask ChatGPT something, ChatGPT answers. The answer will be better or worse depending on the prompt, but it is a direct relationship, the basic sender, message, receiver model of communication we all studied years ago. With Google, the path is much less direct.
Google AI Overviews: why do the links keep you inside Google?
With Google, I do not even need to open Gemini. I go to Google Search, and in most cases it routes me to Gemini through AI Overviews. And it does not end there. Lately (I am writing this in September 2026), AI Overviews has started citing Google itself: click on a link and there is a good chance you land on another Gemini prompt in AI Mode instead of a web page.
On top of that, Google has reworked its interface so the user never has to leave: more answers, new angles, new ways to approach the question, often without citing any source at all, even though almost none of that content was created by Google beyond its own guidelines.
Google is presenting other people’s content as its own, without giving the authors credit. Make of that what you will.
From a UX point of view this is a textbook conflict of interest. The interface is designed to keep people on Google’s pages, not to get them to the best answer. Users notice. So do the people who wrote the answers in the first place.
Does Google pay publishers for AI Overviews?
Partly because of all this, Google has just started testing a way to pay content owners. It is called the “AI contribution pilot”, and it shows payouts in Search Console when a site’s content contributes significantly to AI-generated answers. For now it covers a few dozen small and mid-sized publishers, and Google has not said how contribution is measured or how much anyone gets paid. In SEO and content-creator circles, most people already assume the amounts will be tiny.
Could bad UX end Google’s monopoly?
Google is managing to anger users at every level: professionals, business owners, advertisers, everyday users, academics. ChatGPT’s growth is clear proof of that, and so are the webmaster, SEO and AEO forums, the advertisers, and a legion of users who do get their questions answered, with one small problem. In my daily use, Gemini is clearly behind ChatGPT, and the gap in the number of errors between the two is enormous. In a world where people lean more and more on AI as a crutch, a tool that keeps handing them serious mistakes is something no serious company can allow.
Google’s dominance is undeniable and will stay that way for many years. In 2024 a US federal court even ruled that Google holds an illegal monopoly in search. But this may be the first time in the history of tech that we can watch people stop trusting Google, once the most desirable option there was.
Not because of its logo, its colors or its guidelines. Because of the experience.
“Don’t be evil.” Really.
Frequently asked questions
Partly. In 2015 Alphabet adopted “do the right thing” in its code of conduct. In 2018 Google removed “Don’t be evil” from the preface of its own code of conduct, keeping it only as a closing line. Its principles page still says “You can make money without doing evil.”
No. By the numbers the companies report, ChatGPT and Meta AI both have more users than the Gemini app, even though Google also pushes Gemini through the AI features in Search.
Google has been testing links inside AI Overviews that open AI Mode instead of an external page. The effect is that users stay inside Google’s own AI interface rather than visiting the sites the information came from.
Google is testing it. In September 2026 it detailed an “AI contribution pilot” that shows payouts in Search Console to a small group of publishers, without public details on how contribution is measured or how much is paid.
Google still dominates search. But the fast growth of ChatGPT and growing frustration with AI Overviews among users, publishers and advertisers suggest that trust in Google is eroding for the first time in years.
At Dorve we have been designing experiences for nearly three decades, long before AI was part of the conversation. If your product needs an honest look at its UX, see what we do or tell us what you are building.