Google’s AI Overviews are sending users to DuckDuckGo

🦆 Tired of Google’s AI-first search results? You’re not the only one. An anti-AI search movement is gaining momentum as a growing number of users turn to DuckDuckGo’s No-AI traditional search experience. Following Google’s latest I/O developer event, where the company unveiled its AI-first search engine revamp, privacy-focused search engine DuckDuckGo saw its highest-ever single-day search traffic last Monday.

DuckDuckGo’s no-AI search page is seeing record traffic, with visits up nearly 30% w-o-w and US iPhone installs soaring almost 70%. Now, among the iOS App Store’s top three most downloaded browsers, the company is benefiting from growing AI fatigue. The numbers are hard to ignore — could AI chip away at Google’s long-held search-engine reign?

Why DuckDuckGo?

When Google started prioritizing AI Overviews over its signature blue-link pages, most online users were not on board. DuckDuckGo quickly became an easy alternative, launching a new browser extension for Chrome and Firefox that lets users opt out of AI search results by setting noai.duckduckgo.com as their default search engine. On the no-AI search page, users won’t find AI-assisted answers, chat prompts, or the flood of AI-generated images.

DuckDuckGo is not anti-AI, however. The software company launched its own AI chatbot, Duck.ai, late last year. It has access to many popular models and offers a subscription plan that provides access to the latest models. Long before the AI backlash, DuckDuckGo was known as a privacy-first alternative to Google, given its strong data-protection features. Given that reputation, it’s no surprise that those frustrated with Google’s aggressive AI push are now flocking to the platform.

Why the backlash against Google’s AI Overviews?

AI has been plagued by concerns over accuracy. Even its strongest advocates acknowledge that fact-checking generated answers remains essential — and Google’s AI Overviews is no different. An analysis by AI start-up Oumi reveals that AI Overviews is accurate nine times out of 10. However, given Google’s estimated 5 tn searches per year, even a small margin of error could still result in tens of mns of inaccuracies. Google’s own analysis of Gemini 3, the model powering AI Overviews, found that it produces incorrect information 28% of the time.

Ungrounded answers: AI Overviews have shifted Google’s search results from clearly credited publishers to an aggregated layer of information that draws from multiple sources. The main concern is whether Google’s AI-generated answers are accurately linked to and grounded in the correct sources. Even if the model produces correct responses, the websites it links to more often than not don’t completely support the information provided. Ungrounded correct answers jumped from 37% in October last year to 56% this February. You may get the right answer — but how do you know it’s actually correct?

Facebook and Reddit are among the top-cited sources in Google’s AI Overviews. Yet another analysis by Oumi suggests a striking pattern: Facebook was cited 5% of the time in accurate responses, compared to 7% in inaccurate ones — raising further questions about reliability. Companies are catching on, with some even attempting to game the system by flooding popular platforms like Reddit with branded content in hopes of being picked up in Google’s AI search results.

Another challenge is the variability of Google’s system, which can produce different responses to identical queries submitted at different times, even seconds apart, meaning one answer may be accurate while another is not.

Publishers are bearing the brunt

Beyond concerns over accuracy, AI has also faced criticism for effectively “freeloading” off online publishers and websites, with attribution often vague and web traffic taking a hit.

To combat the issue, UK regulators yesterday demanded that Google set clear attribution and links to publishers’ content in its AI-generated search results. The UK Competition and Markets Authority (CMA) also announced that publishers can now opt out of allowing Google’s AI-generated tools to use their content. This new rule, to which Google already agreed, assures opted-out publishers that they may not be penalized by Google through downranking in general search results.

As traditional blue links give way to AI-generated answers, the need for clear rules and accountability will only grow. Whether Google’s AI-first vision prevails or users continue to favor traditional search experiences like DuckDuckGo, AI is still a technology that requires careful oversight — not just relentless integration.